JP5087724B2 - Image compression apparatus, image expansion apparatus, image compression program, and image expansion program - Google Patents

Description

  The present invention relates to an image compression apparatus, an image expansion apparatus, an image compression program, and an image expansion program.

  2. Description of the Related Art Conventionally, a technique for compressing image data is known in order to facilitate handling of image data (still images and moving images) with a large amount of data. In such a technique, the input image is subjected to DCT (Discrete Cosine Transform), quantized according to human visual characteristics for each frequency band, and further encoded to compress the image data. However, when such a technique is used, the entropy increases when the difference between the pixel values of adjacent pixels in the input image is large. Therefore, the compression rate is not so high. And mosquito noise occurred.

As a technique for solving such a problem, an image data compression method is known in which image data is divided into two and encoded (see, for example, Patent Document 1). In such a technique, pixels of image data are grouped into a pixel in which the pixel value of an adjacent pixel exceeds a predetermined threshold and other pixels, and encoding is performed for each grouped pixel. As a result, the group of pixels that do not exceed the predetermined threshold value has a smaller pixel value entropy, and thus the compression rate can be increased.
JP 2005-294934 A

  However, in the technique described in Patent Document 1, the compression ratio is improved in one of the image data divided into two, but since all the pixels are finally subjected to encoding, the image data As a whole, there was a problem that a very high compression rate could not be realized.

  Further, in the technique described in Patent Document 1, in order to reduce the number of pixels to be encoded, even when only pixels whose pixel values exceed a predetermined threshold are encoded, the pixel values of adjacent pixels Pixels are extracted depending on whether or not the difference value exceeds a predetermined threshold value, it is not possible to accurately extract a characteristic pixel that becomes a change point of the pixel value, and the reproducibility of the image expanded from the compressed data can be reduced. There was a problem of lowering.

  The present invention has been made in view of the above, and is an image compression apparatus, an image expansion apparatus, an image compression program, and an image that have a high compression rate, prevent occurrence of block noise and mosquito noise, and have high image reproducibility. The purpose is to provide a decompression program.

To solve the above problems and achieve the object, the present invention is an image compression apparatus for compressing images data, and the pixel value of a predetermined direction of the pixels of the image data, in the predetermined direction The first pixel among the pixels, the second pixel different from the first pixel, the first pixel, and the second pixel shown in the graph representing the relationship with the position of the pixel A change amount calculating means for calculating an area of a region having a third pixel existing between them as a vertex as a change amount of an increase / decrease rate of a pixel value, and the calculated change amount exceeds a predetermined threshold in, the third pixel, and wherein the pixel and determines the characteristic pixel determination unit is a pixel showing the change point of pixel values in said image data, and the determined pixel value of the feature pixel, and the feature pixel Corresponding to the distance from the feature pixel adjacent to the feature pixel And wherein the pixel storage means for only storing, it is characterized by and an entropy encoding means for entropy encoding the stored the pixel value and the distance to the feature pixel storage means.
According to another aspect of the invention, there is provided a first pixel among the pixels shown in a graph representing a relationship between a pixel value of a pixel in a predetermined direction of image data and a position of the pixel in the predetermined direction; The area of the region having the second pixel different from the first pixel and the third pixel existing between the first pixel and the second pixel as the vertex is the change amount of the increase / decrease rate of the pixel value. When the calculated amount of change exceeds a predetermined threshold, the third pixel is determined as a feature pixel that is a pixel indicating a change point of the pixel value in the image data, and is determined. An image decompression device that decompresses compressed data obtained by entropy-encoding the pixel value of the feature pixel and the distance between the feature pixel and the feature pixel adjacent to the feature pixel, and receiving the compressed data Data receiving means, and the received Entropy decoding is performed on the compressed data to restore the pixel value and the distance of the feature pixel, and the relationship between the restored pixel value of the feature pixel and the position of the feature pixel calculated from the distance. And a pixel value interpolation means for interpolating the pixel values of the pixels between the feature pixels with pixel values on a straight line or a curve passing through the pixel values of the feature pixels in the graph to be represented.

The present invention is also an image compression program for causing a computer to execute an image compression method for compressing image data, wherein a pixel value of a pixel in a predetermined direction of the image data and a pixel position in the predetermined direction are A first pixel among the pixels, a second pixel different from the first pixel, and a first pixel existing between the first pixel and the second pixel, as shown in the graph representing the relationship. A change amount calculating step of calculating an area of a region having the top of 3 pixels as a change amount as a change amount of an increase / decrease rate of the pixel value, and when the calculated change amount exceeds a predetermined threshold, A feature pixel determination step of determining the pixel of the image data as a feature pixel that is a pixel indicating a change point of the pixel value in the image data, the pixel value of the determined feature pixel, and the feature pixel and the feature pixel adjacent to each other The distance to the feature pixel, Characterized in that to execute the entropy coding process for entropy coding the distance the pixel value and stored in the feature pixel storage means for storing in association with, to the computer.
According to another aspect of the invention, there is provided a first pixel among the pixels shown in a graph representing a relationship between a pixel value of a pixel in a predetermined direction of image data and a position of the pixel in the predetermined direction; The area of the region having the second pixel different from the first pixel and the third pixel existing between the first pixel and the second pixel as the vertex is the change amount of the increase / decrease rate of the pixel value. When the calculated amount of change exceeds a predetermined threshold, the third pixel is determined as a feature pixel that is a pixel indicating a change point of the pixel value in the image data, and is determined. An image decompression program for causing a computer to execute an image decompression method for decompressing compressed data obtained by entropy encoding the pixel value of the feature pixel and the distance between the feature pixel and the feature pixel adjacent to the feature pixel. , The compressed data Receiving compressed data receiving step, entropy decoding the received compressed data to restore the pixel value of the feature pixel and the distance, the restored pixel value of the feature pixel, and the A pixel value interpolation step of interpolating the pixel values of the pixels between the feature pixels with a pixel value on a straight line or a curve passing through the pixel values of the feature pixels in a graph representing the relationship between the positions of the feature pixels calculated from the distance; Are executed by a computer.

  According to the present invention, since only the feature pixel determined based on the pixel value and position of the pixel in the predetermined direction is the target of encoding, the number of pixels that are the target of encoding is reduced. The rate can be improved. In addition, since DCT conversion is not performed on the input image, it is possible to prevent occurrence of block noise and mosquito noise. In addition, by determining the feature pixel based on the pixel value and position of the pixel in a predetermined direction, the feature pixel that is a pixel indicating the change point of the pixel value is accurately determined, and the target is to be encoded without omission. Therefore, it is possible to improve image reproducibility when decompressing compressed data.

  Exemplary embodiments of an image compression apparatus, an image expansion apparatus, an image compression program, and an image expansion program according to the present invention are explained in detail below with reference to the accompanying drawings. The present invention is not limited to these embodiments.

(First embodiment)
FIG. 1 is a block diagram showing a configuration of an image compression apparatus 100 according to the present embodiment. In this embodiment, the image data is described as a color moving image. However, the present invention is not limited to this, and a monochrome moving image, a color still image, or a monochrome still image may be used instead of a color moving image.

  The image compression apparatus 100 according to the present embodiment includes an image data receiving unit 101, an image data storage unit 102, a color conversion unit 103, a frame storage unit 104, a change amount calculation unit 105, and a feature pixel determination unit 106. A feature pixel storage unit 107, a quantization unit 108, and an entropy encoding unit 109.

  The image data receiving unit 101 receives color moving image data (RGB pixel values) input from the outside (for example, a digital video camera) and stores the received image data in the image data storage unit 102. The image data storage unit 102 stores the image data received from the image data receiving unit 101.

  The color conversion unit 103 acquires image data of one frame among a plurality of frames (screens) stored in the image data storage unit 102, and converts the acquired image data into respective pixel values of YCrCb. The frame storage unit 104 stores each pixel value of YCrCb converted by the color conversion unit 101 for each frame. In this embodiment, the image data is converted to YCrCb as an example, but it is not limited to YCrCb, and it may be converted to other color expressions YIQ, XYZ, L * a * b *, CMYK, etc. Good. Also, using the characteristic that human vision is less sensitive to CrCb (color difference) component than Y (luminance) component, the pixel value of CrCb is averaged from the pixel values of four pixels to one pixel. Value is used, and image data with a data amount of 1/4 is used.

  The change amount calculation unit 105 uses the pixel value of each YCrCb obtained from the image data stored in the frame storage unit 104 and the pixel position in a predetermined direction (for example, the scanning direction of the image data) to increase or decrease the pixel value. The amount of change is calculated. The change amount of the increase / decrease rate of the pixel value is a scale indicating whether or not the pixel values of the pixels arranged in a predetermined direction are changing at a constant increase / decrease rate.

  As an example, in the graph showing the relationship between the pixel value of each pixel aligned in the scanning direction and the position of the pixel in the scanning direction, the area of the region connecting the three pixels that are subject to determination of the change amount of the increase / decrease rate of the pixel value is A case where the change amount of the increase / decrease rate of the pixel value is described.

  If the pixel values of the three pixels change at a constant rate of increase / decrease, the pixels drawn on the graph are substantially in a straight line, and the area of the region connecting the three pixels is reduced. Therefore, if the area is small, the relationship that the change amount of the increase / decrease rate of the pixel value is small is established. On the other hand, if the pixel values of the three pixels for which the amount of change in the increase / decrease rate of the pixel value is not changed at a constant increase / decrease rate, the pixels drawn on the graph deviate from the straight line. The area of the region that connects is increased. Therefore, if the area is large, the relationship that the change amount of the increase / decrease rate of the pixel value is large is established. In this way, by determining the area of the region connecting the pixel values of the three pixels, it is possible to determine whether the increase / decrease rate of the pixel values of the three pixels has changed.

  The feature pixel determination unit 106 determines whether or not the change amount of the increase / decrease rate of the pixel value calculated by the change amount calculation unit 105 exceeds a threshold value. When the amount of change in the increase / decrease rate of the pixel value exceeds the threshold, the feature pixel determination unit 106 determines that the pixel indicating the point where the increase / decrease rate of the pixel value has changed is a feature pixel, and the pixel value of the feature pixel and the feature The distance between the pixel and the feature pixel adjacent to the feature pixel (hereinafter referred to as the distance between the feature pixels) is stored in the feature pixel storage unit 107. Note that the distance between the feature pixel and the feature pixel adjacent to the feature pixel is obtained as a difference value between the positions of the two feature pixels.

  The above relationship will be further described with reference to FIG. FIG. 2 is a graph showing a part of the relationship between pixel values and pixel positions of pixels arranged in the scanning direction of image data. First, the change amount calculation unit 105 calculates the area connecting the pixels 22 and 23 as the change amount of the increase / decrease rate of the pixel value with the pixel 21 positioned at the head of the pixels arranged in the scanning direction as a feature pixel. As shown in FIG. 2, since the pixels 21 to 23 are drawn on a substantially straight line, the area is almost zero. Therefore, the feature pixel determination unit 106 determines that the amount of change does not exceed the threshold and does not determine the pixel 22 as a feature pixel. Next, an area connecting the pixel 21, the pixel 23, and the pixel 24 is calculated. As shown in FIG. 2, since the pixel 21, the pixel 23, and the pixel 24 are also drawn on a substantially straight line, the area is almost zero. Therefore, the feature pixel determination unit 106 determines that the amount of change does not exceed the threshold and does not determine the pixel 23 as a feature pixel. Next, an area 28 (horizontal line region) of a region connecting the pixel 21, the pixel 24, and the pixel 25 is calculated. Regarding the pixel 21, pixel 24, and pixel 25, the rate of increase of the pixel value is slightly changed, but is not so large, and thus the area of the region connecting them is calculated to be smaller than the threshold value. Thereby, the feature pixel determination unit 106 determines that the amount of change does not exceed the threshold, and determines that the pixel 24 is not a feature pixel. As described above, information on the pixels 22, 23, and 24 that are not determined as feature pixels (specifically, pixel values and distances of the respective pixels) is not stored in the feature pixel storage unit 107, and the amount of data is reduced. Thus, the compression rate is improved.

  Next, the case where the pixel 21, the pixel 25, and the pixel 26 are the target pixels will be described. The change amount calculation unit 105 calculates an area 29 (vertical line region) of a region connecting the pixel 21, the pixel 25, and the pixel 26 as a change amount of the increase / decrease rate of the pixel value. In this case, since the increase / decrease rate of the pixel value from the pixel 21 to the pixel 25 and the increase / decrease rate of the pixel value from the pixel 25 to the pixel 26 are greatly different, the area of the region connecting the pixel 21, the pixel 25, and the pixel 26 is larger than the threshold value. A large value is calculated. Accordingly, the feature pixel determination unit 106 determines that the amount of change exceeds the threshold, and determines that the pixel 25 existing between the pixel 21 and the pixel 26 is a feature pixel. Accordingly, the pixel value of the feature pixel 25 and the distance L between the feature pixel 21 and the feature pixel 25 are stored in the feature pixel storage unit 107. As a result, when decompressing the compressed data, the image data is restored using the pixel value of the feature pixel 25 that is the change point of the pixel value, so that the reproducibility of the image is improved.

  Next, the change amount calculation unit 105 calculates an area connecting the pixel 25, which is a feature pixel, the pixel 26, and the pixel 27, and the feature pixel determination unit 106 determines the feature pixel. In this way, by sequentially determining the feature pixels, the pixel values of the feature pixels and the pixel values of the pixels arranged in the scanning direction can be represented by the distance between the feature pixels. Note that the direction in which the three pixels for determining the characteristic pixel are selected need not be limited to the scanning direction described above, and three pixels may be selected from the reverse direction. Alternatively, an arbitrary pixel in the scanning direction may be selected, and that pixel may be selected as a first feature pixel, and three pixels may be sequentially selected in both directions to determine the feature pixel.

  Further, the case of other image data will be described. FIG. 3 is a graph showing a part of the relationship between pixel values and positions of pixels arranged in the scanning direction in other image data. The graph shown in FIG. 3 is assumed to be image data including characters and graphics. Of the plurality of pixels between the pixels 31 to 34, the area of the region connecting the three selected pixels is calculated to be almost 0 and does not exceed the threshold value, so the pixels 32 to 33 are not determined as feature pixels. . Thereby, the amount of data can be greatly reduced, and the compression rate is improved.

  On the other hand, the area of the pixel 31, the pixel 34, and the pixel 35 is calculated as a value that exceeds the threshold value, so that the pixel 34 is determined as a feature pixel. In this way, by holding the edge portion where the brightness and color of characters and figures greatly change as the feature pixel, the edge portion can be accurately reproduced when decompressing the compressed data. Since the change amount of the increase / decrease rate of the pixel value is the area, even if the difference value between the pixel value of the pixel 34 and the pixel value of the pixel 35 is not so large, the area depends on the distance between the pixels. Since it becomes large, it is possible to reliably grasp the change point of the pixel value.

  The feature pixel storage unit 107 stores the pixel values of the feature pixels determined one after another by the feature pixel determination unit 106 and the distance between the feature pixels. FIG. 4 is an explanatory diagram illustrating an example of a data configuration of the feature pixel storage unit 107. The feature pixel storage unit 107 stores, for each YCrCb, the pixel value of the feature pixel in each pixel row arranged in the scanning direction of the image data and the distance between the feature pixels in association with each other.

  The quantization unit 108 quantizes the pixel value of the feature pixel stored in the feature pixel storage unit 107. All pixel values are quantized using the characteristics of the human eye that cannot recognize minute changes in pixel values. This quantization is realized, for example, by setting the pixel value to 1/2 or 1/4. As a result, it is possible to improve the data compression rate while suppressing the deterioration of the image quality recognized visually.

  The entropy encoding unit 109 entropy encodes the quantized pixel value of the feature pixel and the distance between the feature pixels. More specifically, the entropy encoding unit 109 performs entropy encoding on the quantized pixel values of the feature pixels and the distance between the feature pixels to perform Huffman code, arithmetic code, LZW (Lempel Ziv Welch) code, PPM (Prediction by Partial compressed data is generated by expressing it with a code or the like.

  Next, image compression processing of the image compression apparatus 100 configured as described above will be described. 5A and 5B are flowcharts illustrating the image compression processing procedure performed by the image compression apparatus 100.

  First, the image data receiving unit 101 receives image data (step S501), and stores the received image data in the image data storage unit 102 (step S502). The color conversion unit 103 converts the RGB image data into YCrCb image data (step S503). The color conversion unit 103 stores the converted YCrCb image data in the frame storage unit 104 (step S504). The frame storage unit 104 stores a Y component pixel value, a Cr component pixel value, and a Cb component pixel value for each frame.

  The change amount calculation unit 105 acquires the pixel values of one row of pixels arranged in the image data scanning direction from the frame storage unit 104 (step S505). The change amount calculation unit 105 calculates the area of the region connecting the three pixels by using the pixel values of the three pixels and the position in the scanning direction (step S506). Here, the position of the pixel can be indicated by the number of the pixel from the end of the image data in the pixels arranged in the scanning direction. The feature pixel determination unit 106 determines whether or not the calculated area exceeds a predetermined threshold (step S507).

  When it is determined that the calculated area exceeds a predetermined threshold (step S507: Yes), the feature pixel determination unit 106 determines the pixel value of the pixel determined as the feature pixel and the distance between the feature pixels as the feature pixel. The data is stored in the storage unit 107 (step S508). When it is determined that the calculated area does not exceed a predetermined threshold (step S507: No), the process proceeds to step S509.

  The feature pixel determination unit 106 determines whether or not the processing for all the rows has been completed (step S509). If it is determined that all the processing for one line has been completed (step S509: Yes), the feature pixel determination unit 106 determines whether all the processing for one frame has been completed (step S510). On the other hand, if it is determined that the processing for all the lines has not been completed (step S509: No), the process returns to step S506.

  If it is determined that all the processing for one frame has been completed (step S510: Yes), the quantization unit 108 acquires the pixel value of the feature pixel of one frame and the distance between the feature pixels from the feature pixel storage unit 107 ( Step S511). If it is determined that all the processing for one frame has not been completed (step S510: No), the process returns to step S505, and the pixel values of the pixels in the next row are acquired.

  The quantization unit 108 quantizes the pixel value of the feature pixel in one frame (step S512). The entropy encoding unit 109 entropy encodes the quantized pixel value of the feature pixel and the distance between the feature pixels to generate compressed data (step S513). Note that color moving image compression data is generated by sequentially executing the above-described processing for all of a plurality of image data.

  In this way, by setting the image data as the pixel value of the feature pixel and the distance between the feature pixels, the image data is represented by data obtained by approximating the pixel value with a straight line for each pixel row, and the pixel data between the feature pixels is Since it can be omitted, the amount of image data is reduced and the compression rate of the image data is improved. In addition, by holding the pixel values of feature pixels whose pixel value increase / decrease rate changes, the image can be restored using pixels indicating a characteristic image such as an edge portion. improves.

  In addition, the area obtained as the change amount of the increase / decrease rate of the pixel value when determining the feature pixel can be calculated by integrating the pixel value. Therefore, the processing load can be reduced compared to the case of using multiplication or division in DCT processing.

  In the above-described embodiment, the area of the region connecting the three pixels is the amount of change in the increase / decrease rate of the pixel value, but is not limited to the area, and may be a value indicating the amount of change in the increase / decrease rate of the pixel value. Any may be used. FIG. 6 is a graph showing a part of the relationship between the pixel values of the pixels arranged in the scanning direction of the image data and the pixel positions. As shown in FIG. 6, among the three pixels 41 to 43, the pixel 42 that is a candidate for the feature pixel, a straight line extending from the pixel 42 in the pixel value axis direction, and the other two pixels 41 and 43 are connected. The distance from the intersection 44 with the straight line may be the amount of change.

  Next, an image expansion apparatus that expands the compressed data generated by the image compression apparatus 100 described above will be described. FIG. 7 is a block diagram showing the configuration of the image expansion apparatus 200 according to this embodiment. The image expansion apparatus 200 includes a compressed data receiving unit 201, a compressed data storage unit 202, an entropy decoding unit 203, an inverse quantization unit 204, a feature pixel storage unit 205, a pixel value interpolation unit 206, image data And a storage unit 207.

  The compressed data receiving unit 201 receives compressed data transmitted from the outside. The outside is, for example, a server (hereinafter referred to as a content server) that is connected to a network such as the Internet or a LAN (Local Area Network), stores content, and distributes or transmits content. The compressed data receiving unit 201 stores the compressed data received from the outside in the compressed data storage unit 202. The compressed data storage unit 202 stores the compressed data received by the compressed data receiving unit 201.

  The entropy decoding unit 203 performs entropy decoding on the compressed data stored in the compressed data storage unit 202, and restores the quantized pixel values of the feature pixels and the distances between the feature pixels. The inverse quantization unit 204 inversely quantizes the quantized pixel value of the feature pixel, and restores the pixel value of the feature pixel and the distance between the feature pixels. The inverse quantization unit 204 stores the pixel value of the restored feature pixel and the distance between the feature pixels in the feature pixel storage unit 205. The feature pixel storage unit 205 stores the pixel value of the restored feature pixel and the distance between the feature pixels.

  The pixel value interpolation unit 206 interpolates the pixel value of the pixel between the feature pixels from the pixel value of the feature pixel stored in the feature pixel storage unit 205 and the distance between the feature pixels, and obtains image data including the pixel values of each YCrCb. Restore. More specifically, the pixel value interpolation unit 206 interpolates the pixel values of the pixels between the feature pixels with values on a straight line connecting the pixel values of the feature pixels. Hereinafter, a case where the pixel values of the pixels between the characteristic pixels are interpolated with a straight line will be described with reference to FIG.

  FIG. 8 is an explanatory diagram illustrating an example in which pixel values of pixels between feature pixels are interpolated with a straight line. In FIG. 8, the characteristic pixels are pixels 51 to 53. Since it can be determined from the distance between the pixel 51 and the pixel 52 that the three pixels between the pixel 51 and the pixel 52 are omitted, the straight line connecting the feature pixel 51 and the feature pixel 52, the pixel 51 and the pixel 52 Intersections 54 to 56 with straight lines indicating the positions of the pixels in between are obtained, and pixel values Y11, Y12, and Y13 of the omitted pixels are calculated. In this way, the pixel values of the omitted pixels can be interpolated as pixel values on a straight line connecting feature pixels.

  Further, the pixel value interpolation unit 206 may interpolate the pixel values of the omitted pixels by connecting the pixel values of the feature pixels with a curve instead of a straight line and calculating the pixel values on the curve. FIG. 9 is an explanatory diagram illustrating an example of interpolating pixel values of pixels between feature pixels with a curve. 9, the pixel values and positions of the feature pixels are the same as those in FIG. 8, and the feature pixels are the pixels 51 to 53. The pixel value interpolation unit 206 generates a curve (for example, a spline curve) connecting the pixels 51 to 53, obtains intersections 57 to 59 between the curve and a straight line indicating the position of the pixel between the pixels 51 and 52, Pixel values Y21, Y22, and Y23 of the omitted pixels are calculated. In this way, by using the values on the curve connecting the characteristic pixels as pixel values, it is possible to interpolate with pixel values closer to the original image data, thereby improving image reproducibility.

  Further, the pixel value interpolation unit 206 stores the pixel values of the pixels between the feature pixels interpolated by the pixel value interpolation unit 206 and the pixel values of the feature pixels in the order of the pixels arranged in the image data. The image data storage unit 207 stores the decompressed image data.

  Next, the image expansion process of the image expansion apparatus 200 configured as described above will be described. FIGS. 10A and 10B are flowcharts illustrating the image expansion processing procedure performed by the image expansion apparatus 200. FIG.

  First, the compressed data receiving unit 201 receives the compressed data (step S1001), and stores the received compressed data in the compressed data storage unit 202 (step S1002). The entropy decoding unit 203 decodes the compressed data stored in the compressed data storage unit 202 (step S1003). When the entropy decoding unit 203 decodes the compressed data, the pixel value of the quantized feature pixel and the distance between the feature pixels are obtained. The inverse quantization unit 204 inversely quantizes the pixel value of the quantized feature pixel (step S1004). Thereby, the pixel value of the feature pixel and the distance between the feature pixels are restored.

  The inverse quantization unit 204 stores the restored pixel value and distance of the feature pixel in the feature pixel storage unit 205 (step S1005). The feature pixel storage unit 205 stores the pixel value and distance of the Y-component feature pixel for each frame, the pixel value and distance of the Cr-component feature pixel, and the pixel value and distance of the Cb-component feature pixel. Next, the pixel value interpolation unit 206 acquires the pixel values and distances of the feature pixels for one line of the image data from the feature pixel storage unit 205 (S1006).

  The pixel value interpolation unit 206 interpolates the pixel values of the pixels omitted between the feature pixels based on the pixel values of the feature pixels and the distance between the feature pixels, and restores the image data (step S1007). Specifically, as described above, the pixel value is calculated by obtaining the intersection of the straight line connecting the characteristic pixels and the straight line indicating the position of the pixel to be interpolated, and the image data is restored from the calculated pixel value. As described above, the image decompression apparatus 200 does not perform inverse DCT, so that the processing load when decompressing the compressed data does not increase, and therefore, the image decompression apparatus 200 can be applied to an apparatus that does not have high processing capability.

  The pixel value interpolation unit 206 stores the interpolated pixel value and the pixel value of the feature pixel in the image data storage unit 207 in the order of the pixels arranged in the image data (step S1008). The pixel value interpolation unit 206 determines whether or not processing for one row has been completed (step S1009). If it is determined that the processing for one row has been completed (step S1009: Yes), the pixel value interpolation unit 206 determines whether all the processing for one frame has been completed (step S1010). On the other hand, if it is determined that all the lines have not been processed (step S1009: No), the process returns to step S1007.

  If it is determined that all the processes for one frame have been completed (step S1010: Yes), the process ends. If it is determined that all the processes for one frame have not been completed (step S1010: No), the process returns to step S1006 to obtain the pixel values of the next row of pixels.

  As described above, the image compression apparatus 100 generates compressed data in which pixel values of pixels between feature pixels are omitted, and transmits the compressed data to the image expansion apparatus 200, thereby reducing the amount of data received by the image expansion apparatus 200. Therefore, the reception time can be shortened. Further, since the processing load when decompressing the compressed data is small, it is possible to restore the image data in a short processing time. Accordingly, when the image expansion apparatus 200 includes a monitor that displays the expanded image data, power consumption can be reduced, and thus the image data can be displayed on the monitor for a long time.

  In addition, since the image compression apparatus 100 and the image expansion apparatus 200 according to the present embodiment do not use DCT or inverse DCT, JPEG (Joint Photographic Experts Group) format data or MPEG (Moving Picture Experts Group) format data is used. Generation of mosquito noise and block noise around the edge as seen when stretched can be prevented.

  As a modification, in the image compression processing described above, instead of determining feature pixels for all pixel rows arranged in the scanning direction of image data, feature pixels are determined for each predetermined region of the image data. Then, the pixel value of the determined feature pixel and the distance between the feature pixels may be quantized and entropy coded. Here, the predetermined area specifically refers to an area obtained by dividing image data by a predetermined number of pixels (for example, 8 pixels), or for each predetermined block (for example, for every 8 pixels × 8 pixels). This is an area obtained by dividing image data. As described above, the characteristic pixels are determined for the pixels arranged in the scanning direction in a predetermined pixel width or a predetermined block unit, and the pixel values of the characteristic pixels are held, whereby the dispersion of the pixel values is reduced, and the entropy. Since the code length can be shortened, the compression rate can be increased.

  Further, the predetermined area does not need to be a rectangle, and is an area composed of polygons or curves previously indicated by coordinates on the image data, an area determined as a person using a known technique, a person's face, and the like. It may be a determined region, a region including a region determined to be a person or a human face, and the like. In addition, the above-described image compression is performed for image data in a predetermined area, and a known image compression technique (for example, other JPEG or JPEG2000) is used for image data in other areas. In addition, it is possible to obtain a high-quality image in an area where it is desired to prevent image quality deterioration and to increase the compression rate of the image data.

  Furthermore, another modified example will be described. First, similarly to the above-described modification, feature pixels are determined with a width of a predetermined number of pixels in the scanning direction (for example, with a width of 8 pixels) in one frame of image data. In this case, since the first one of the eight pixels is always determined as a feature pixel as a reference, the feature pixel is determined for these pixels, that is, pixels arranged in a direction orthogonal to the scanning direction. The pixel value of the feature pixel determined in this way and the distance between the feature pixels are quantized and entropy encoded. As a result, the feature pixels are also determined in the direction orthogonal to the scan direction, compared to the case where the feature pixels are determined with a predetermined number of pixels in the scan direction (every 8 pixels). Therefore, the compression rate of image data can be increased.

  Further, the column of pixels for obtaining the feature pixel need not be limited to the above-described rectangular end of the 8-pixel width. For example, the above-described polygon or a region determined to be a person is also arranged on the outer periphery of the region. Among the pixels, the characteristic pixel may be determined with respect to a pixel that is first detected by scanning from the scanning direction. Since the pixels in the outer periphery of such an area are reference pixels, they are determined as feature pixels. However, by determining the feature pixels from these pixels, the compression target can be reduced, so the compression rate of image data can be reduced. Can be increased.

  Further, the above-described image compression apparatus 100 and image expansion apparatus 200 may be applied to a monitoring system including a monitoring camera and a security device. The surveillance camera determines the feature pixel from the captured image data, quantizes the pixel value of the determined feature pixel, and encodes the quantized feature pixel value and the distance between the feature pixels to generate compressed data To do. A transmission unit included in the monitoring camera transmits the generated compressed data to a monitoring device connected via a network. The monitoring device decodes and inverse-quantizes the received compressed data, and generates image data by interpolating the pixel value from the restored pixel value of the feature pixel and the distance between the feature pixels. The monitoring device displays the generated image data on a monitoring monitor monitored by a guard.

  Also, content (image data of programs such as news and sports) broadcasted from broadcast stations by digital terrestrial television broadcasting for mobile devices to mobile phones, and content (image data such as movies) distributed from content providers' content servers ) May be applied to the image compression apparatus 100 and the image expansion apparatus 200 described above.

Further, the above-described image compression apparatus 100 and image expansion apparatus 200 are combined with CT (Computed
You may apply to the imaging device in medical fields, such as Tomography) and MRI (magnetic resonance imaging).

  Further, the image compression apparatus 100 according to the present embodiment is not limited to the above-described apparatus and method, and the CPU (processor) in the image compression apparatus 100 performs image compression for realizing the above-described embodiment. You may implement | achieve by operating various devices by reading a program from ROM (Read Only Memory) and executing it. An image compression program executed by the image compression apparatus 100 includes a module configuration including the above-described units (image data reception unit, color conversion unit, change amount calculation unit, feature pixel determination unit, quantization unit, entropy encoding unit). As actual hardware, the CPU reads the image compression program from the ROM and executes it, so that the above-described units are loaded onto the main storage device, and the image data receiving unit 101, the color conversion unit 103, the change amount calculation The unit 105, the feature pixel determination unit 106, the quantization unit 108, and the entropy encoding unit 109 are generated on the main storage device.

  Note that the image data storage unit 102, the frame storage unit 104, and the feature pixel storage unit 107 are configured on the main storage device, as well as all commonly used devices such as an HDD (Hard Disk Drive), an optical disk, and a memory card. It can also be configured by a storage medium.

  Similarly, the CPU in the image expansion apparatus 200 of the present embodiment is realized by operating various devices by reading out and executing an image expansion program for realizing the above-described embodiment from the ROM. Also good. The image expansion program executed by the image expansion apparatus 200 has a module configuration including the above-described units (compressed data reception unit, entropy decoding unit, inverse quantization unit, pixel value interpolation unit), and actual hardware As the CPU reads out and executes the image compression program from the ROM, the above-described units are loaded onto the main storage device, and the compressed data receiving unit 201, the entropy decoding unit 203, the inverse quantization unit 204, and the pixel value interpolation unit 206 is generated on the main memory. Note that the image compression program and the image expansion program are provided by being incorporated in advance in a ROM or the like. In addition, the compressed data storage unit 202, the feature pixel storage unit 205, and the image data storage unit 207 are configured on the main storage device, and are configured by all commonly used storage media such as an HDD, an optical disc, and a memory card. You can also

  The image compression program and the image expansion program are files that can be installed or executed, and can be read by a computer such as a CD-ROM, a flexible disk (FD), a CD-R, or a DVD (Digital Versatile Disk). You may provide by recording on a recording medium.

  Further, the image compression program and the image expansion program may be stored on a computer connected to a network such as the Internet and provided by being downloaded via the network. Further, the image compression program and the image expansion program may be provided or distributed via a network such as the Internet.

(Second Embodiment)
A second embodiment will be described with reference to the accompanying drawings. As for the image compression apparatus and the image expansion apparatus to which the present embodiment is applied, the differences from the first embodiment will be described. For the same parts as in the first embodiment, refer to the description of the first embodiment.

  FIG. 11 is a block diagram showing the configuration of the image compression apparatus 300 according to this embodiment. An image compression apparatus 300 according to the present embodiment includes an image data receiving unit 101, an image data storage unit 102, a color conversion unit 103, a frame storage unit 104, a change amount calculation unit 105, and a feature pixel determination unit 106. A feature pixel storage unit 107, a difference value calculation unit 310, a difference value storage unit 311, a quantization unit 108, and an entropy encoding unit 109. Note that portions having the same functions as those in the first embodiment are denoted by the same reference numerals, and description thereof is omitted here.

  The difference value calculation unit 310 calculates the pixel value of the feature pixel in one row and the position in the scanning direction of the image data stored in the feature pixel storage unit 107, the pixel value of the feature pixel in the next row, and the position in the scanning direction. The difference value of the pixel value and the difference value of the position in the scanning direction between the feature pixels are calculated. The difference value storage unit 311 stores the calculated difference value of the pixel value for each row and the difference value of the position in the scanning direction.

  FIG. 12 is a graph showing the relationship between the pixel values of feature pixels in two adjacent rows in one image data and the position in the scanning direction. The curves shown in FIG. 12 depict pixel values of pixels in two rows in the image data scanning direction, and black circles indicate feature pixels. The feature pixels in one row are pixels 61 to 65, and the feature pixels in the next row are pixels 71 to 75. For each feature pixel, the difference value of the pixel value of the corresponding pixel and the difference value of the position in the scanning direction are calculated. For example, the pixel value “Yi, j” of the feature pixel 62 and the position “Xi, j” in the scanning direction, the pixel value “Yi, j + 1” of the feature pixel 72, and the position “Xi, j + 1” in the scanning direction. ", The difference value ΔY of the pixel value of the feature pixel 72 and the difference value ΔX of the position in the scanning direction can be obtained by the following equations. Note that i indicates the order in which the feature pixels in one row are determined, and j indicates what number row in the image data.

ΔY = Yi, j + 1−Yi, j (1)
ΔX = Xi, j + 1−Xi, j (2)

  ΔY indicated by the expression (1) is a difference value between pixel values, and thus can be either positive or negative. When the difference value of the pixel value is 0, flag information with the difference value of 0 may be stored in the difference value storage unit 311 to reduce the data amount. In addition, ΔX indicated by the equation (2) is a position difference value, and thus takes either a positive or negative value. When the position difference value is 0, flag information having a difference value of 0 may be stored in the difference value storage unit 311 to reduce the data amount.

  The difference value calculation unit 310 calculates the difference value between the pixel values of the feature pixels from the previous row and the difference value between the positions in the scanning direction. The quantization unit 108 quantizes the calculated pixel value difference value and the position difference value in the scanning direction, and the entropy encoding unit 109 performs entropy encoding to generate compressed data.

  In the image expansion apparatus 200, the compressed data receiving unit 201 receives the compressed data, the entropy decoding unit 203 entropy decodes the compressed data, and the inverse quantization unit 204 performs inverse quantization. Next, the pixel value interpolation unit 206 restores the pixel value of the feature pixel and the distance between the feature pixels from the restored difference value of the pixel value of the feature pixel and the difference value of the position in the scanning direction. Further, the pixel value interpolation unit 206 interpolates the pixel values of the pixels between the feature pixels from the pixel values of the feature pixels and the distance between the feature pixels, and restores the image data. As described above, by using the difference value of the pixel value or the difference value of the position in the scanning direction, the entropy is reduced, and the compression rate is improved.

  As a modification, instead of feature pixels in different rows in the scanning direction of the same image data, pixel value difference values and position difference values in the scanning direction in the same row in a plurality of image data May be used to generate compressed data.

  As another modification, instead of obtaining the difference value of the position of the feature pixel instead of the difference value of the position of the pixel in the scanning direction, the compressed data is obtained using the position value of the feature pixel itself and the difference value of the pixel value. You may create it. Note that the position of the feature pixel can be indicated by the number of the pixel from the end of the image data in the pixels arranged in the scanning direction, as in the above-described embodiment.

(Third embodiment)
A third embodiment will be described with reference to the accompanying drawings. As for the image compression apparatus and the image expansion apparatus to which the present embodiment is applied, the differences from the first embodiment will be described. For the same parts as in the first embodiment, refer to the description of the first embodiment.

  The feature pixel determination unit 106 stores a plurality of threshold values, and determines whether or not the target pixel is a feature pixel from each of the plurality of threshold values and the change amount calculated by the change amount calculation unit 105. . Here, a value corresponding to the level at which image data is compressed is set for each of the plurality of threshold values. For example, three levels of feature pixels can be determined by using three threshold values. The feature pixel determination unit 106 stores the pixel value of the feature pixel and the distance between the feature pixels in the feature pixel storage unit 107 for each threshold value. The feature pixel storage unit 107 stores the pixel value of the feature pixel and the distance between the feature pixels in association with each level. The quantization unit 108 quantizes the data for each level, and the entropy encoding unit 109 performs entropy encoding on the data for each level. As a result, compressed data in which pixel values of feature pixels corresponding to a plurality of levels and distances between feature pixels are encoded from one image data is generated.

  FIG. 13 is a graph showing a part of the relationship between pixel values and pixel positions of pixels arranged in the scanning direction of image data. The curve shown in FIG. 13 is a drawing of pixel values of pixels arranged in the scanning direction of the image data, and the characteristic pixels are indicated by black circles, black triangles, and black squares. For example, when level 1 is the level with the highest compression rate, the threshold value for determining the level 1 feature pixel is the largest value. In FIG. 13, pixels 81, 85, and 89 indicated by black circles are determined as level 1 feature pixels.

  Next, if level 2 is the second highest compression level, the threshold value for determining level 2 is the second largest value, and pixels 81, 83, 85, 87, 89 indicated by black circles and black triangles are displayed. It is determined as a feature pixel. When the feature pixel determination unit 106 stores the pixel value of the feature pixel and the distance between the feature pixels in the feature pixel storage unit 107, if the feature pixels determined with different threshold values overlap, the feature pixel determination unit 106 determines with a larger threshold value. The feature pixel determined, that is, the feature pixel determined at level 1 is deleted. Specifically, pixels 83 and 87 are stored as level 2 feature pixels.

  Furthermore, if level 3 is the level with the lowest compression rate, the threshold value for determining level 3 is the smallest value, and pixels 81 to 89 indicated by black circles, black triangles, and black squares are determined as feature pixels. When the feature pixel determination unit 106 stores the pixel value of the feature pixel and the distance between the feature pixels in the feature pixel storage unit 107, if the feature pixels determined with different threshold values overlap, the feature pixel determination unit 106 determines with a larger threshold value. The feature pixels determined, that is, the feature pixels determined at level 1 and level 2 are deleted. Specifically, pixels 82, 84, 86, and 88 are stored as level 3 feature pixels.

  As described above, the image compression apparatus determines feature pixels from a single image data using a threshold corresponding to the level, and generates compressed data for each level, so that various communication bands and processing of the image expansion apparatus can be performed. The compressed data corresponding to the capability and the size of the monitor for displaying the image can be easily generated by combining one or more levels of compressed data.

  As a modification, a case where compressed data is transmitted in response to a compressed data transmission request from the image expansion apparatus will be described. The image compression apparatus generates compressed data having different levels in advance, and stores the generated compressed data for each level in the storage unit. The location where compressed data is stored need not be limited to the image compression apparatus, and may be stored in a content server or the like different from the image compression apparatus. In response to an image data transmission request including the type of the image expansion device transmitted from the image expansion device, the image compression device transmits compressed data at a level corresponding to the type of the image expansion device from the compressed data stored in the storage unit. To do.

  In this way, in response to a request from the image decompression apparatus, one or a plurality of levels of compressed data corresponding to the request among the compressed data generated from one image data and stored in advance for each level are transmitted. Thus, compressed data corresponding to the type of the image expansion device can be transmitted to the image expansion device without increasing the communication load between the image compression device and the image expansion device. That is, compressed data generated from one image data for each level is transmitted, for example, if the transmission destination is a mobile phone, only level 1 compressed data is transmitted, and if the transmission destination is HDTV (High Definition television) By transmitting 1 to 3 compressed data, it is possible to transmit compressed data corresponding to various requests from high compression to lossless with one compressed data stored at a plurality of levels.

  On the other hand, for example, when generating level 2 image data, the image expansion apparatus receives level 1 compressed data and level 2 compressed data. Next, each compressed data is entropy-decoded and dequantized to restore the pixel value of the feature pixel and the distance between the feature pixels for each level. Next, the pixel value of the feature pixel for each level and the distance between the feature pixels are sorted, and the pixel value of the feature pixel for each row and the distance between the feature pixels are generated. Thereafter, the pixel value interpolation process described above is executed to generate image data. As described above, the image decompression apparatus according to the present embodiment can restore image data from compressed data with the same function and configuration regardless of the level of image data decompression. Standardization of the expansion device can be achieved.

  In the above-described embodiment, the case has been described in which feature pixels that do not overlap for each level are retained. However, all feature pixels determined by the threshold value for each level may be retained in an overlapping manner.

  As mentioned above, although this invention was demonstrated using the 1st-3rd embodiment, a various change or improvement can be added to the said embodiment. The functions and configurations described in the first to third embodiments can be freely combined.

It is a block diagram which shows the structure of the image compression apparatus concerning this Embodiment. It is a graph which shows a part of relationship between the pixel value of the pixel arranged in the scanning direction of image data, and the position of a pixel. It is a graph which shows a part of relationship of the pixel value and position of the pixel located in a line with the scanning direction in other image data. It is explanatory drawing which shows an example of a data structure of a feature pixel memory | storage part. It is a flowchart which shows the image compression process sequence which an image compression apparatus performs. It is a flowchart which shows the image compression process sequence which an image compression apparatus performs. It is a graph which shows a part of relationship between the pixel value of the pixel arranged in the scanning direction of image data, and the position of a pixel. It is a block diagram which shows the structure of the image expansion | extension apparatus concerning this Embodiment. It is explanatory drawing which shows an example which interpolates the pixel value of the pixel between characteristic pixels with a straight line. It is explanatory drawing which shows an example which interpolates the pixel value of the pixel between characteristic pixels with a curve. It is a flowchart which shows the image expansion process procedure which an image expansion apparatus performs. It is a flowchart which shows the image expansion process procedure which an image expansion apparatus performs. It is a block diagram which shows the structure of the image compression apparatus in this Embodiment. It is a graph which shows the relationship between the pixel value of the feature pixel of two adjacent rows in one image data, and the position in a scanning direction. It is a graph which shows a part of relationship between the pixel value of the pixel arranged in the scanning direction of image data, and the position of a pixel.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 300 Image compression apparatus 101 Image data receiving part 102 Image data memory | storage part 103 Color conversion part 104 Frame memory | storage part 105 Change amount calculation part 106 Feature pixel determination part 107 Feature pixel memory | storage part 108 Quantization part 109 Entropy encoding part 200 Image decompression | decompression Device 201 Compressed data receiving unit 202 Compressed data storage unit 203 Entropy decoding unit 204 Inverse quantization unit 205 Feature pixel storage unit 206 Pixel value interpolation unit 207 Image data storage unit 310 Difference value calculation unit 311 Difference value storage unit

Claims (10)

An image compression apparatus for compressing images data,
A first pixel of the pixels, the first pixel, and a graph representing a relationship between a pixel value of a pixel in a predetermined direction of the image data and a position of the pixel in the predetermined direction; A change amount for calculating an area of a region having apexes of different second pixels and the third pixel existing between the first pixel and the second pixel as a change amount of a pixel value increase / decrease rate A calculation means;
When the calculated amount of change exceeds a predetermined threshold, said third pixel, and determines the characteristic pixel determination unit, wherein the pixel is a pixel showing the change point of pixel values in said image data,
Feature pixel storage means for storing the determined pixel value of the feature pixel and the distance between the feature pixel and the feature pixel adjacent to the feature pixel in association with each other;
Entropy encoding means for entropy encoding the pixel value and the distance stored in the feature pixel storage means;
An image compression apparatus comprising: The first pixel of the pixels shown in the graph showing the relationship between the pixel value of the pixel in the predetermined direction and the position of the pixel in the predetermined direction, and a second different from the first pixel An intersection of a straight line connecting the first pixel and a straight line extending in a pixel value axis direction from a third pixel existing between the first pixel and the second pixel, and the obtained intersection and the A change amount calculating means for calculating a distance from the third pixel as a change amount of an increase / decrease rate of the pixel value;
The image compression according to claim 1, wherein the feature pixel determination unit determines that the third pixel is the feature pixel when the calculated amount of change exceeds a predetermined threshold. apparatus. The change amount calculating means calculates the change amount from the first pixel, the second pixel, and the third pixel included in a predetermined region in the image data,
The image compression apparatus according to claim 2 , wherein the entropy encoding unit performs entropy encoding on the pixel value and the distance stored in the feature pixel storage unit for each of the predetermined regions. The change amount calculating unit further includes the first pixel, the second pixel, and the first pixel among the pixels arranged on the outer periphery in the predetermined region and detected first by scanning from the predetermined direction. Calculating the amount of change from a third pixel;
The entropy encoding means further entropy coding the pixel value and the distance of characteristic pixels of the feature pixel direction perpendicular to the predetermined direction stored in the storage means, to claim 3, characterized in The image compression apparatus described. A difference value of pixel values between the feature pixels of the first row in the predetermined direction and the feature pixels of the second row adjacent to the first row; the feature pixels of the first row; Difference value calculating means for calculating a difference value of the position in the predetermined direction with the feature pixel in the second row;
Difference value storage means for storing the calculated difference value of the pixel value and the difference value of the position in association with each other; and
The entropy encoding means, any one of claims 1 to 4, characterized in that, the entropy coding the difference value of the difference value and the position of the pixel value stored in the difference value storage means The image compression apparatus described in 1. And the feature pixel of the predetermined row of the first image data among the plurality of image data, the pixel value of the feature pixel of the predetermined row of the second image data to be continuous with the first image data A difference value calculating means for calculating a difference value and a difference value between positions of the feature pixels of the first image data and the feature pixels of the second image data in the predetermined direction;
Difference value storage means for storing the calculated difference value of the pixel value and the difference value of the position in association with each other; and
The entropy encoding means, any one of claims 1 to 4, characterized in that, the entropy coding the difference value of the difference value and the position of the pixel value stored in the difference value storage means The image compression apparatus described in 1. The feature pixel determining means determines the feature pixel using thresholds corresponding to a plurality of levels,
The feature pixel storage means stores the pixel value of the feature pixel and the distance between the feature pixel and the feature pixel adjacent to the feature pixel in association with each level,
The entropy encoding means, each said level, either the pixel value of the feature pixel stored in the feature pixel storage means, of claims 1 to 6, characterized in that, to entropy encoding and said distance The image compression apparatus as described in any one. The first pixel of the pixels shown in the graph showing the relationship between the pixel value of the pixel in the predetermined direction of the image data and the position of the pixel in the predetermined direction is different from the first pixel The area of a region having the second pixel and the third pixel existing between the first pixel and the second pixel as a vertex is calculated as the change amount of the increase / decrease rate of the pixel value. When the amount of change exceeds a predetermined threshold, the third pixel is determined as a feature pixel that is a pixel indicating a change point of a pixel value in the image data, and the determined feature pixel An image expansion device that expands compressed data obtained by entropy encoding a pixel value and a distance between the feature pixel and a feature pixel adjacent to the feature pixel,
Compressed data receiving means for receiving the compressed data;
Entropy decoding means for entropy decoding the received compressed data and restoring the pixel value of the feature pixel and the distance;
In the graph showing the relationship between the restored pixel value of the feature pixel and the position of the feature pixel calculated from the distance, the feature pixel is a pixel value on a straight line or a curve passing through the pixel value of the feature pixel. A pixel value interpolating means for interpolating the pixel values of the pixels between,
An image expansion apparatus comprising: An image compression program for executing an image compression method for compressing images data to the computer,
A first pixel of the pixels, the first pixel, and a graph representing a relationship between a pixel value of a pixel in a predetermined direction of the image data and a position of the pixel in the predetermined direction; A change amount for calculating an area of a region having apexes of different second pixels and the third pixel existing between the first pixel and the second pixel as a change amount of a pixel value increase / decrease rate A calculation process;
When the change amount calculated exceeds a predetermined threshold, said third pixel, and the characteristic pixel determination step of determining said pixel is a pixel showing the change point of pixel values in said image data,
The pixel value and the distance stored in the feature pixel storage means for storing the determined pixel value of the feature pixel and the distance between the feature pixel and the feature pixel adjacent to the feature pixel in association with each other are entropy. An entropy encoding step for encoding;
A computer-executable image compression program. The first pixel of the pixels shown in the graph showing the relationship between the pixel value of the pixel in the predetermined direction of the image data and the position of the pixel in the predetermined direction is different from the first pixel The area of a region having the second pixel and the third pixel existing between the first pixel and the second pixel as a vertex is calculated as the change amount of the increase / decrease rate of the pixel value. When the amount of change exceeds a predetermined threshold, the third pixel is determined as a feature pixel that is a pixel indicating a change point of a pixel value in the image data, and the determined feature pixel An image decompression program for causing a computer to execute an image decompression method for decompressing compressed data obtained by entropy encoding a pixel value and a distance between the feature pixel and a feature pixel adjacent to the feature pixel,
A compressed data receiving step for receiving the compressed data;
Entropy decoding the received compressed data, and restoring the pixel value and the distance of the feature pixel; and
A pixel value on a straight line or a curve passing through the pixel value of the feature pixel in the graph representing the relationship between the restored pixel value of the feature pixel and the position of the feature pixel calculated from the distance, and between the feature pixels A pixel value interpolation step for interpolating the pixel values of the pixels;
An image decompression program that causes a computer to execute the program.

Images