JP3018713B2 - Image communication device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image communication apparatus for compressing / expanding a still image and having a function of displaying a still image on a small screen.

[0002]

2. Description of the Related Art As a method for displaying a still image on a monitor, there are a main screen display for displaying still screen image data on the entire monitor as shown in FIG. 2 and a sub-screen display for reducing and displaying the above-mentioned main screen display. Can be divided. Generally, child screen image data is created by processing parent screen image data.

FIG. 3 is a configuration diagram of a conventional general image communication apparatus. In FIG. 3, a still image taken in by a still image input unit 1 composed of a scanner, a camera or the like is sent to an encoding unit 2 as input image data, and is encoded. The compressed image data compressed and encoded by the encoding means 2 is stored in the code memory 3. The decoded image data decoded by the compressed image data decoding means 4 output from the code memory 3 and restored by the decoding means 4 is displayed on the monitor 5. The line interface 6 can transmit and receive the compressed image data stored in the code memory 3 to and from a communication network. Reference numeral 7 denotes an overall control unit that controls the entire image communication apparatus.

[0004] In addition to the above configuration, by providing multiplexing means, not shown here, it is also possible to perform communication of voice, moving surface, data, and the like.

Next, a method of creating a restored main screen image by compressing / expanding still image data and processing the processed image to create sub-screen image data will be described.

FIGS. 4A and 4B are block diagrams of the still picture encoding means 2 and the still picture decoding means 4 which are generally used for compressing and expanding a color still picture. It is.

First, the encoding process will be described. FIG.
In (a), 8 performs discrete cosine transformation on image data taken in from the still image input means 1 as described later,
Discrete cosine transform means (hereinafter, referred to as “DCT means”) for outputting the coefficients of the discrete cosine function in block units, and 9 refers to the quantization table for encoding 10 for the coefficients of the discrete cosine function output from the DCT means 8. This is a quantization means for performing quantization. The coding quantization table 10 used here is set by the quantization table setting means 11 according to a command from the overall control unit 7 for controlling the entire image communication apparatus. Huffman encoding means 12 performs entropy encoding of the quantized value output from the quantization means 9 and outputs compressed image data to the code memory 3.

[0008] Here, Huffman coding means that a code as short as possible is assigned to a code data sequence in the order of the highest probability of occurrence, so that the transformed code data sequence after the DCT transform and quantization process is converted into a code amount per pixel. To approach the lower limit of the average information (entropy),
Of the various techniques for creating this code, the easiest to implement in hardware using tables,
Commonly used.

Next, in the decoding process, a process reverse to the encoding process is performed. That is, the compressed image data extracted from the code memory 3 is restored through the Huffman decoding unit 13, the inverse quantization unit 14, and the inverse DCT unit 16. Generally, the same table as the encoding quantization table 10 is used for the decoding quantization table 15 that provides the quantization table used by the inverse quantization means 14, and
It is set by the overall control unit 7 and the quantization table setting unit 11. The capacity of the compressed image data and the image quality of the restored image data generally depend on the quantization table value of the decoding quantization table 15.

[0010] Discrete cosine transform means that when a signal having high correlation between pixels is viewed on a frequency axis, signal power is intensively distributed to low frequency components. A process of encoding and compressing the amount of information. The coefficient of the function obtained by the discrete cosine transform is a function of the frequency. The Fourier transform is best known as a means for orthogonal transform into the frequency domain.However, since the Fourier transform has a complex structure including a complex component in the function coefficient, the power concentration degree to the low frequency component is currently present. Is close to the optimum value, and a DCT transform using only a cosine function that is easy to implement as an LSI with a high-speed operation algorithm is in use.

FIG. 5 is a diagram showing input / output coefficients used for discrete cosine transform. As described above, encoding and decoding of image data generally uses not only the horizontal pixel correlation but also the two-dimensional DCT transform using the vertical direction. Is performed in block units. Numerals 1 to 64 shown in FIG. 5 are obtained by arranging coefficients of a function obtained by performing discrete cosine transform of image data in order from a low frequency, and an upper left corner of a coefficient block usually has an average value (DC coefficient). ), And arrange them so that the higher the frequency goes toward the lower right corner. When these DCT-transformed coefficients are further quantized, information can be compressed because the lower right of the coefficient block, that is, most of high-frequency components become 0.

[0012] Due to the characteristics of the human eye, the lower the frequency component, the sharper the visual sensation.
Usually, high frequency components are unnecessary, and it is considered that high frequency components can be removed by the above-described method.

Next, a description will be given of a conventional means for creating child screen image data from the parent screen image data created by the above-described method.

FIGS. 6A and 6B show a conventional small-screen image data creating method. In FIG. 6A, reference numeral 17 denotes a thinning means, which is the simplest method of thinning out the image data so that the main screen image data becomes an arbitrary reduced screen to create the child screen data. Since the information of the image data is completely discarded, the image becomes a particularly coarse small-screen image.

In FIG. 6B, reference numeral 18 denotes digital filter means comprising a plurality of filters. This method uses a multi-stage digital filter that creates sub-screen image data with reference to a plurality of image data before, after, left and right of the target image data, and creates optimal sub-screen image data. And an additional circuit and buffer are required.

[0016]

As described above, in the conventional configuration, when the main screen image data is simply decimated, the image becomes coarse, and a circuit which requires an operation is required to improve the image quality. Leads to higher prices,
There were pros and cons.

An object of the present invention is to solve the above-mentioned problem and to provide a simple small-screen creating means for creating small-screen image data of good image quality by using a generally used still image compressing / expanding means. It is an object.

[0018]

SUMMARY OF THE INVENTION In order to achieve the above object, the present invention quantizes a spectrum signal obtained by linearly transforming main screen image data and arranges the quantized spectrum signals as coefficient values in a matrix. Converting means for obtaining a restored image data having a resolution different from that of the main screen image data from a signal in a low frequency region in the coefficient value, a quantization table previously set by a restored image table setting means. And decoding means for extracting one or a plurality of pixels of the image data in the low frequency region from the decoding means to obtain restored image data.

[0019]

According to the above arrangement, in the encoding means, the original image signal is converted into a spectrum signal by linear conversion and output as a compressed image signal quantized by a quantization table. The compressed image signal is inversely quantized by a decoding unit using a preset table different from that of the encoding unit, and only a low-frequency component is extracted to obtain a small-screen image signal.

[0020]

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a configuration diagram showing a decoding process in one embodiment of the present invention. In the figure, reference numeral 19 denotes a decoding means, which is a Huffman decoding means 20, an inverse quantization means 21, an inverse DCT means 22, and a decoding quantization means 23.
, A main screen quantization table setting unit 24, and a small screen quantization table setting unit 25. The Huffman decoding unit 20, the inverse quantization unit and the inverse DCT unit 22 are the same as those shown in the prior art. is there. The same table as that at the time of encoding, which is designated by the overall control unit 26 and set by the main screen quantization table setting unit 24, is used for the decoding quantization table 23. Reference numeral 27 denotes a thinning unit that extracts only a low-frequency component within a certain range from the image data decoded by the decoding unit 19 and creates small-screen image data.

At the time of normal parent screen restoration image data creation,
As described above, the compressed image data is transmitted to the Huffman decoding unit 2.
0, the main screen restored image data is created via the inverse quantization means 21 and the inverse DCT means 22.

When the small-screen image data is created, the quantization table for the small screen is set in the decoding quantization table 23 by the overall control unit 26 via the small-screen quantization table setting means 25, and the inverse quantization means is set. In step 21, the image data processed with reference to the quantization table is converted into inverse DCT means 22.
By performing inverse DCT processing as shown in FIG. 5 and extracting only one or a plurality of pixels of the image data in which the low-frequency component has been decoded by the thinning means 27, clear small-screen image data can be obtained.

For example, in order to input only the DC component of the DCT coefficient (pixel 1 in FIG. 5) to the inverse DCT means 21, a value 0 is input to the pixels 2 to 64 in the inverse DCT means 22. Decoding inverse quantization value into quantization table for decoding 2
Setting to 3 makes it possible. By inputting only the DC component to the inverse DCT means 22, the 64 restored image data of 8 pixels × 8 pixels becomes the DC component of the 64 restored main screen image data with reference to the normal main screen quantization table. Only the restored image is obtained, and a small-screen image having a resolution of 1/8 can be created. Also, inverse DCT means 1
Similar processing is possible even when some AC coefficients other than the DC coefficient are input to 6 and used.

The small-screen image obtained in this manner is an image that has been passed through a low-pass filter that incorporates information of 64 pieces of pixel data as compared with an image obtained by simply thinning out the main screen image data. Excellent image.

[0025]

As described above, by quantizing the original image signal and extracting only the low frequency components, it is possible to obtain small-screen image data of a clear image.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an image communication apparatus according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram of a screen display in a conventional image communication device.

FIG. 3 is a block diagram showing a configuration of the image communication apparatus.

FIG. 4A is a configuration diagram of an encoding unit of the image communication device. FIG. 4B is a configuration diagram of a decoding unit of the image communication device.

FIG. 5 is a configuration diagram of input / output coefficients of the image communication apparatus.

6A is a diagram illustrating a thinning unit of the image communication device. FIG. 6B is a diagram illustrating a digital filter unit of the image communication device.

[Explanation of symbols]

 Reference Signs List 19 decoding means 20 Huffman decoding means 21 inverse quantization means 22 inverse DCT means 23 decoding quantization table 24 parent screen quantization table setting means 25 child screen quantization table setting means 26 overall control unit 27 thinning means

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification code FI H04N 7/30 H04N 7/133 Z

Claims (1)

(57) [Claims] 1. Huffman encoding means for performing entropy encoding of compressed image data, parent screen quantization table setting means having a parent screen quantization pattern, and child screen quantization table having a child screen quantization pattern. When decoding the main screen image data, a quantization pattern is extracted and output from the main screen quantization table setting means, and when decoding the sub screen image data, the quantization pattern is output from the sub screen quantization table setting means. Quantization table for extracting and outputting the quantization pattern, and inverse quantization for inversely quantizing the compressed image data output from the Huffman encoding means based on the quantization pattern obtained from the quantization table for decoding. Means, and DC output from the inverse quantization means.
Inverse DCT means for performing inverse DCT processing on the T signal and outputting the same as restored image data; and, when creating small-screen image data, arranging the restored image data output from the inverse DCT means in a matrix as coefficient values. An image communication apparatus comprising: thinning means for obtaining restored image data having a resolution different from that of an image original signal by using one or a plurality of coefficient values in a low frequency region among the coefficient values.