JP5691609B2 - Film mode conversion device, film mode conversion method, and program - Google Patents

Description

  The present invention relates to a film mode detection device, a film mode detection method, and a program, and in particular, a film mode detection for displaying an interlaced scanning telecine video signal on a progressive scanning display such as a liquid crystal TV. The present invention relates to an apparatus, a film mode detection method, and a program.

  In recent years, progressive (sequential scanning) displays such as liquid crystal displays and plasma displays have been used in televisions and personal computers in place of cathode ray tubes.

  In these progressive scan displays, a progressive method of 24 frames per second capable of displaying an image with little flicker is employed.

  On the other hand, in television broadcasting, the NTSC system, which is an interlace (interlaced scanning) system of 30 frames (60 fields) per second, is employed. Therefore, in a television or personal computer equipped with a progressive scan display, an IP conversion process (Interlace to Progressive) for converting an interlace (interlaced) video to a progressive (sequential scan) video of 24 fields per second may be performed. It is done.

  However, since cinema images are originally 24 frames per second (sequential scanning) images, 24 frames per second are required to display cinema images on a television with a liquid crystal display or plasma display, which is a progressive scanning display. IP conversion processing (Interlace to Progressive) that receives telecine video converted to interlaced (interlaced scanning) video at 30 frames per second (interlaced scanning) and converts this telecine video into progressive (sequential scanning) video Display in line.

  Specifically, in the process of converting the above-described cinema frame at 24 frames per second to interlaced video at 30 frames (60 fields) per second, this cinema video is generated by 23 pull-down conversion processing to create an interlaced telecine video.

  Further, the process of converting the telecine video into the original progressive video creates a progressive video by reverse 23 pull-down conversion processing of the telecine video.

  In order to perform the inverse 23 pulldown conversion processing, 23 pulldown sequence detection is performed.

  FIG. 8 is an explanatory diagram (part 1) illustrating 23 pull-down sequence detection. FIG. 8 shows a 23 pull-down conversion process for converting (pull-down) the film material 11 into a video image 12.

The film material 11 is a progressive (sequential scanning) image of 24 frames per second, and the video image 12 is an interlaced (interlaced scanning) image of 30 frames (60 fields) per second. Image A film material 11 is divided into A ₁ and A ₂ of the video image 12, constituting the first frame to fit them together constitute the first and second fields respectively.

Image B of the film material 11 is not only divided into _{B 1} and _{B 2} of the video image 12, the repeat field is copied to the _{B 3.}

  That is, each frame of the film material 11 can be converted into 2 fields by converting odd frames into 2 fields and even frames with 3 fields. Therefore, 24 frames can be converted into 60 fields.

  FIG. 9 is an explanatory diagram (part 2) illustrating 23 pull-down sequence detection.

  In FIG. 9, the upper part shows 23 pull-down conversion processing from cinema video to telecine video. The cinema image is a progressive (sequential scanning) system image of 24 frames per second, and the 23 pull-down conversion process repeats the image of one frame every two and two fields before. This is called a repeat field. By adding this repeat field, it is possible to convert a cinema image of 24 frames per second into a telecine image of 60 fields per second.

  In FIG. 9, in the lower part, in order to broadcast the telecine video on the television, the telecine video is subjected to inverse 2-3 conversion processing, and the progressive video is displayed by detecting the 23 pull-down sequence. The telecine video is an interlace video of 60 fields per second, and the 23 pull-down sequence detection detects one repeat field in five frames and restores the original progressive video.

  As described above, since the detection of the repeat field in the 23 pull-down sequence detection is very important in generating a high-quality video by the IP conversion process, various techniques for accurately detecting the repeat field are disclosed. Has been.

  For example, there is a prior art document relating to control for increasing the repeat field detection accuracy as 23 pull-down detection by switching the determination condition according to the presence or absence of noise from the difference in inter-frame difference distribution.

  Further, in the 23 pull-down sequence detection, there is a prior art document relating to control for performing detection using the fact that the distribution of the absolute value of the interframe difference is biased to near 0 in a cycle of 5 fields.

Japanese Patent Laid-Open No. 2008-98938 JP 2007-300152 A

  However, the background art described above has the following problems.

  When telecine video is detected from interlaced video that contains a lot of noise from old cinema video or noise caused by encoding, repeat field detection is performed. Since the gradation difference is often greater than or equal to the threshold value, the repeat field image that should be the repeat field is determined as the other field, and is detected and deteriorated. That is, since the distribution of the absolute value of the difference between frames is dispersed by noise, there is a problem that the resistance to noise is low in principle.

  Accordingly, the present invention has been made in view of the above problems, and an object thereof is to provide a film mode detection device, a film mode detection method, and a program that are resistant to block noise and old cinema images.

One aspect of an apparatus for solving the above-described problem is that an inter-frame difference absolute value calculating unit that inputs an input interlaced video signal for a plurality of continuous fields and calculates an inter-frame difference absolute value, and the video A difference absolute value determination unit that compares the absolute value between frames and a motion threshold for each pixel of a signal, and the absolute value between frames for a pixel for which the absolute value between frames is determined to be greater than the motion threshold. And a luminance gradient value obtained by differentiating the luminance between the frames of the peripheral pixels of the pixel with respect to a pixel for which the absolute difference between frames is determined to be larger than the luminance change threshold. Is greater than or equal to the luminance gradient threshold value in the comparison determination with the luminance gradient threshold value, the pixel is an edge detection determination unit that determines the vertical and oblique edges of the moving subject, and one line A moving pixel cumulative count unit that counts the cumulative number of moving pixels as many as the number of diagonal edges found, a moving pixel cumulative count number and a still pixel cumulative count number in the current field, An approximation degree determination unit that detects a repeat field compared to the still pixel cumulative count number and performs an inverse 23 conversion process to return to the original progressive video;
have.

  As described above, according to the disclosed technology, the repeat field is detected by paying attention to the vertical and oblique edge portions where the change of the moving pixel of the video signal is remarkable, so that the film is resistant to block noise and old cinema video. A mode detection device, a film mode detection method, and a program can be provided.

It is a block diagram of the film mode detection apparatus in one Embodiment of this invention. It is a whole flowchart of the field mode detection procedure in one Embodiment of this invention. It is explanatory drawing (the 1) of edge detection determination by this invention. It is explanatory drawing (the 2) of edge detection determination by this invention. It is explanatory drawing of the approximation degree determination by this invention. It is a detailed flowchart (the 1) of the field mode detection procedure in one Embodiment of this invention. It is a whole flowchart (the 2) of the field mode detection procedure in one Embodiment of this invention. It is explanatory drawing (the 1) which shows 23 pulldown sequence detection. It is explanatory drawing (the 2) which shows 23 pulldown sequence detection.

FIG. 1 is a configuration diagram of a film mode detection apparatus according to an embodiment of the present invention. As shown in FIG. 1, the film mode detection apparatus includes two field memories 1 ₁ , 1 ₂ , an inter-frame difference absolute value calculation unit 2, a difference absolute value determination unit 3, a still pixel cumulative number counting unit 4, A bias detection unit 5, a luminance determination unit 6, an edge detection determination unit 7, a moving pixel cumulative number counting unit 8, and an approximation degree determination unit 9 are provided.

The field memories 1 ₁ and 1 ₂ store F (n) and F (n−1), which are input images of interlaced video for two consecutive fields, respectively.

The inter-frame difference absolute value calculation unit 2 receives F (n) and F (n−1) as video signals output from the field memories 1 ₁ and 1 ₂ and calculates the absolute value of the inter-frame difference.

  The difference absolute value determination unit 3 has the motion of the input video or the motion of the input video depending on whether the difference absolute value calculated by the inter-frame difference absolute value calculation unit 2 is larger or smaller than the motion threshold for each pixel of the video signal. Judge that there is no.

  When the absolute difference between frames is equal to or less than the motion threshold, the still pixel cumulative number counting unit 4 counts up the still pixel cumulative number by 1 assuming that there is no motion for the pixels of the video signal.

  The bias detection unit 5 compares the still pixel cumulative number of the current field and the past number field, and detects a field having a large still pixel cumulative number as a biased image.

  When the difference absolute value determination unit 3 determines that there is a motion of the input signal, the luminance determination unit 6 compares and determines the interframe difference absolute value with the luminance change thresholds 1 and 2 for viewing the luminance change.

  When the luminance determination unit 6 determines that the luminance change is large, the edge detection determination unit 7 compares the luminance gradient value obtained from the differentiation of the inter-frame luminance of the surrounding pixels of the target pixel with the luminance gradient threshold value. When the pixel is equal to or greater than the luminance gradient threshold value in FIG.

  The moving pixel cumulative number counting unit 8 counts up the vertical and diagonal edges that are valid on one line as the moving pixel cumulative number. In other words, since the vertical edge and the diagonal edge always cross the line on the interlace, the moving pixel can be reliably specified.

  The degree-of-approximation determination unit 9 detects the repeat field based on the AND condition of both the determination of the cumulative number of motion pixels in the current field and the determination of the cumulative number of still pixels, and reverse 23 for returning to the original progressive video. A 23 pull-down sequence which is a conversion process is detected.

  With the above configuration, the film mode detection device detects the presence of vertical and oblique edges of a subject moving in units of pixels by comparing the absolute value of the difference between successive frames of the input video and the motion threshold value for viewing the luminance change. Then, the moving pixels in the line direction specified by the vertical and diagonal edges are reliably detected.

  On the other hand, since the repeat field detection necessary for performing the 23 pull-down sequence is performed by detecting a dummy pixel having no motion, the film mode detection device of the present invention has the number of motion pixels and the number of still pixels for the past five fields. It is possible to detect the presence of a repeat field by calculating an approximate value based on the fluctuations of.

  FIG. 2 is an overall flowchart of a field mode detection procedure according to an embodiment of the present invention. The detection procedure of the field mode detection device will be described below.

  S1. The interframe difference absolute value calculation unit 2 calculates the absolute value of the interframe difference of the input video for two consecutive fields.

  S2. The absolute difference determination unit 3 compares the absolute value of the interframe difference calculated by the interframe difference absolute value calculation unit 2 for each pixel with the motion threshold value for determining whether there is a change in luminance, or has motion or no motion. Determine.

  S3. In S2 described above, when the interframe difference absolute value is larger than the motion threshold value (with motion), the luminance determination unit 6 performs a comparison determination with the luminance change threshold value to see the luminance change.

S4. In S3 described above, if the interframe difference absolute value is larger than the luminance change threshold, the luminance gradient value obtained from the differentiation of the interpixel luminance of the surrounding pixels is compared with the luminance gradient threshold for the target pixel. If the luminance gradient threshold value is greater than or equal to the threshold value, the pixel is determined to be an oblique edge or vertical edge of a moving pixel. (This will be explained in detail in FIGS. 3 and 4)
S5. If the absolute difference value is smaller than the motion threshold value in S3 described above, the still pixel cumulative number counting unit 4 counts up by 1 as no motion.

  S6. The current pixel and the past number field still pixel accumulation number are compared, and a field having a large still pixel accumulation number is detected as a biased image.

S7.1: Determine whether the cumulative number of moving pixels counted up by the number of edges found in the line approximates between fields, and determine the cumulative number of still pixels obtained when detecting bias in S6 described above Is detected based on the AND condition with the determination result of whether or not the field is approximated, and a 23 pull-down sequence, which is an inverse 23 conversion process for returning to the original progressive video, is detected. (This will be explained in detail in FIG. 5)
As described above, in the present embodiment, the possibility of the presence of an edge in the target pixel is determined, and then the presence of vertical and oblique edges of the subject moving in pixel units is detected.

  Since the vertical and oblique edges always cross the line on the interlace, the moving pixels are surely detected.

  In S4 described above, the vertical / diagonal edge valid candidates are determined by the number of valid edges (moving pixel count) in the Top field and the number of valid edges (moving pixel count) in the Bottom field for each line. 3 and 4 will be described in detail.

  On the other hand, since the repeat field necessary for detecting the 23 pull-down sequence is a dummy pixel having no motion, the disclosed technique is based on the fluctuation state of the number of motion pixels and the number of still pixels in the adjacent field for the past five fields. The presence of a repeat field is obtained by determining a certain approximate value, which will be described in detail with reference to FIG.

  FIG. 3 is an explanatory diagram (part 1) of edge detection determination according to the present invention. The operation of edge detection determination will be described below.

  S11. As an example, a luminance change determination process between Bot # 4 (Bottom field 4) which is the current field and Bot # 2 (Bottom field 2) two fields before is shown. Specifically, when the absolute difference between frames is equal to or greater than the luminance change threshold in the comparison determination with the luminance change threshold for viewing the luminance change, the possibility that the target pixel has an edge is determined.

  S12. For the target pixel I including the peripheral pixels, the inter-frame luminance differentiation of the peripheral pixels is obtained to obtain a luminance gradient (p0 to p8). Edge candidates can be detected when p4 (luminance gradient) is equal to or greater than the gradient threshold in the comparison determination with the gradient threshold. The Sobel Filter is a filter that detects only edges of vertical lines and horizontal lines, and indicates a weight value (operator) for vertical direction detection.

  S13.1 Vertical and diagonal edge candidates are detected on the line, and the cumulative number of moving pixels is counted up at that time.

  FIG. 4 is an explanatory diagram (part 2) of edge detection determination according to the present invention. The operation of edge detection determination will be described below.

  The determination of the vertical and diagonal edge valid candidates is performed by comparing the edge valid flag 1 for each line of the Top field and the Bottom field.

  That is, when the effective edge number on the Top field side and the effective edge line number on the Bottom field side are equal, and the motion pixel count number on the Top field side is close to the motion pixel count number on the Bottom field side (the number of valid flags exists) In this case, the edge of the Top field and the edge of the Bottom field are the same video.

  The Top field is an even sequence of interlace signals, and the Bottom field is an odd sequence of interlace signals.

  FIG. 5 is an explanatory diagram of the approximation determination according to the present invention. The statistic storage table is provided in the approximation determination unit 9 and detects a repeat field based on data stored in the statistics storage table for determination of the approximation.

  In this storage example, the past five fields are counted by using the motion pixel cumulative number and the still pixel cumulative number of each field between the current field # 5 and the field # 0 five fields before. Adjacent pixel count values are compared.

  The reason why it is only necessary to check the pixel counts for the past five fields is that there are repeat fields used for interlaced video signals every five fields.

The judgment condition of the moving pixel cumulative number is
(1) Is there a field that approximates the cumulative number of motion pixels?
→ Field # 1 and field # 2 are approximated, and field # 3 and field # 4 are approximated.
(2) Which field has the minimum number of moving pixels?
→ Field # 0 and field # 5 are both minimum values.

The criteria for determining the cumulative number of still pixels are:
(3) Which field has the maximum number of still pixels accumulated?
→ Field # 0 and field # 5 are both minimum values.

A field that satisfies the AND conditions (1), (2), and (3) above is selected.
→ Recognizing that field # 0 and field # 5 are repeat fields, as a result, 23 pull-down sequences can be detected. Thus, it is possible to return to the original progressive video.

  FIG. 6 is a detailed flowchart (part 1) of the field mode detection procedure according to the embodiment of the present invention. The detection procedure of the field mode detection device will be described below.

  S21. The interframe difference absolute value calculation unit 2 calculates the absolute value of the interframe difference of the input video for two consecutive fields.

  S22. The difference absolute value determination unit 3 compares the inter-frame difference absolute value and the motion threshold value for each pixel to determine whether there is motion or no motion.

  S23. When there is a motion in S22 described above, a luminance change determination process is performed between the current field and the past field two fields before. Specifically, it is determined whether or not the absolute value of the inter-frame difference is equal to or greater than the luminance change threshold in the comparison determination with the luminance change threshold for viewing the luminance change.

  S24. In S23 described above, when the luminance threshold value is the absolute value of the difference between frames> the luminance change threshold value, the edge detection determination unit 7 detects the presence of vertical and oblique edges of the moving subject.

  S25. If the absolute difference value is smaller than the motion threshold, the still pixel cumulative number counting unit 4 counts up by 1 as no motion.

  S26. The bias detector 5 compares the still pixel cumulative numbers of the current field and the past number fields, and detects a field having a large still pixel cumulative number as a biased image.

  S27. It is determined whether or not the luminance gradient value obtained from the differentiation of the inter-frame luminance of the surrounding pixels is greater than or equal to the luminance gradient threshold in the comparison determination with the luminance gradient threshold.

  S28. When the luminance gradient value is larger than the luminance gradient threshold value in S27 described above, vertical and diagonal edge candidates are detected on one line, and the moving pixel accumulation number counting unit 8 at that time detects the moving pixel accumulation. The number is counted, and the edge valid flag = 1 (valid).

  S29.1: It is determined whether there is no vertical or oblique edge candidate on the line and the processing for one line has been completed. If the processing for one line has not been completed yet, the process proceeds to S21 described above. Return.

  S30. In the above-described S28, when the processing for one line is completed, it is further determined whether or not there are vertical and oblique edge candidates with a certain number of lines or more.

  S31. In the above-described S29, if there are no vertical and oblique edge candidates with a certain number of lines or more, the moving pixel cumulative number is cleared to zero.

  S32. In S29 described above, when there are vertical and diagonal edge candidates with a certain number of lines or more, it is determined whether or not the processing for 1 Field has been completed, and the processing for 1 Field has not been completed yet. The process returns to S21 described above.

  The above detailed flowchart (part 1) details the relationship between the edge valid flag based on the edge detection determination and the count of the cumulative number of motion pixels in the overall flowchart described with reference to FIG. 2, and the edge detection for each line. To the end of one field.

  That is, if only vertical and diagonal edges are seen on the interlace, the lines always cross the line, so there is little blurring of the cumulative number of moving pixels, and if the continuity of the number of lines above a certain level is seen, vertical edge candidates are effective ( If the edge valid flag = 1), attention is paid to the fact that the vertical edge is not erroneously detected as block noise.

  Therefore, it is possible to detect a film mode having resistance against block noise and old cinema video noise.

  FIG. 7 is a detailed flowchart (part 2) of the field mode detection procedure according to the embodiment of the present invention. The detection procedure of the field mode detection device will be described below.

  S33. In FIG. 6, after the end of one screen processing, comparison is made for each line of the determined edge valid flag = 1 (valid), and it is determined whether or not the top edge valid line number is equal to the bottom edge valid line number. Check.

  S34. If the Top side edge effective line number is the same as the Bottom side edge effective line number and the Top side motion pixel count number is close to the Bottom side motion pixel count number in S33 described above, the edge of the Top field Assuming that the edge of the Bottom field is the same video, it is set as a progressive candidate.

  S35. The approximate value determination process is performed by calculating an approximate value from the accumulated number of still pixels and the accumulated number of motion pixels for each field stored in a statistic storage table (see FIG. 5) attached to the approximate degree determination unit 9. Perform the determination process.

  S36. The presence of a repeat field is determined by comparing the motion pixel cumulative number of the current field with the fluctuations of the motion pixel cumulative number for the past five fields.

  S37. In this example, the current field is set to Bottom # 5, the accumulated motion pixel numbers of Bottom # 1 and Top # 2 are approximated, and whether the accumulated motion pixel numbers of Bottom # 3 and Top # 4 are approximated or not is determined. judge.

  S38. It is determined whether or not the cumulative number of motion pixels of Top # 0 and Bottom # 5 is a minimum value from the cumulative number of motion pixels of Bottom 1 to Top # 4.

  S39. The presence of the repeat field is determined by comparing the number of still pixels accumulated in the current field with the variation in the number of still pixels accumulated for the past five fields.

  S40. It is determined whether the cumulative number of still pixels of Top # 0 and Bottom # 5 is the maximum value together with the cumulative number of still pixels of Bottom 1 to Top # 4.

  S41. In S39 described above, when the cumulative number of still pixels of Top # 0 and Bottom # 5 is both the maximum value of the cumulative number of still pixels of Bottom1 to Top # 4, Top # 0 and Bottom # 5 are repeat fields. Recognizing this, transform IP conversion is performed to convert an interlaced video signal into a progressive video signal by IP (Interlace to Progressive).

  S42. If all of the above S37, S38, and S40 are NO (No), the field image is divided into blocks, and block matching is performed within a predetermined search area, thereby performing motion vector detection for each block. In accordance with the motion vector, motion adaptive IP conversion is performed in which an interlace to video signal is converted from an interlace video signal to a progressive video signal.

  Since transform IP conversion and motion adaptive IP conversion are both existing technologies, detailed description thereof is omitted.

  The detailed flowchart (No. 2) described above details the determination of the degree of approximation in the overall flowchart described with reference to FIG. 2. As a result of the determination up to S40, the target repeat field can be detected. Next, the flow up to transform IP conversion in which the repeat field is deleted and IP (Interlace to Progressive) conversion of the video signal of the interlace video into the video signal of the progressive video is shown.

  According to the film mode detection method of the present invention, the repeat field detection necessary for performing the 23 pull-down sequence detection is performed by calculating an approximate value based on fluctuations in the number of moving pixels and the number of still pixels for the past five fields. The presence of can be detected.

The following additional notes are further disclosed with respect to the embodiment including the above examples.
(Appendix 1)
An inter-frame difference absolute value calculation unit that inputs the input interlaced video signals for a plurality of continuous fields and calculates an inter-frame difference absolute value;
A difference absolute value determination unit that compares the difference absolute value between frames and a motion threshold value in units of pixels of the video signal;
A luminance determination unit that compares the absolute difference between frames with a luminance change threshold for a pixel for which the absolute difference between frames is determined to be greater than a motion threshold;
For a pixel for which the interframe difference absolute value is determined to be greater than the luminance change threshold, the luminance gradient value obtained from the differentiation of the interframe luminance of the surrounding pixels of the pixel is the luminance gradient in the comparison determination with the luminance gradient threshold. An edge detection determination unit that determines that the pixel is a vertical or oblique edge of a moving subject when the pixel is equal to or greater than a threshold;
A moving pixel cumulative number counting unit for counting up the cumulative number of moving pixels by the number of vertical and diagonal edges found on one line;
In order to detect the repeat field by comparing the moving pixel cumulative count number and still pixel cumulative count number of the current field with the moving pixel cumulative count number and still pixel cumulative count number for the past five fields, and to return to the original progressive video A degree-of-approximation determination unit that performs an inverse 23 conversion process;
A film mode detection apparatus comprising:
(Appendix 2)
In Appendix 1,
The film mode detection apparatus, wherein the difference absolute value determination unit can be replaced with the luminance determination unit.
(Appendix 3)
In Appendix 1,
The approximation degree determination unit determines the cumulative number of moving pixels and the cumulative number of still images of the current field and the adjacent fields for the past five fields,
In the determination of the cumulative number of motion pixels, the cumulative number of motion pixels in the first field and the second field are approximated, the cumulative number of motion pixels in the third field and the fourth field are approximated, and the current field and the previous five fields are determined. Both the cumulative number of motion pixels in the field No. 1 indicate the minimum value compared with the cumulative number of motion pixels in the first to fifth fields,
In the determination of the still pixel cumulative number, the still pixel cumulative number of the current field and the previous five fields is the maximum value compared with the still pixel cumulative number of the first to fifth fields,
A film mode detecting apparatus, wherein a current field is detected as a repeat field based on an AND condition of the result of moving image cumulative number determination and still pixel cumulative determination result.
(Appendix 4)
A first step of inputting an input interlaced video signal for a plurality of consecutive fields and calculating an inter-frame difference absolute value;
A second step of comparing the absolute value of the interframe difference with a motion threshold for each pixel of the video signal;
A third step of comparing the inter-frame difference absolute value with a luminance change threshold for a pixel for which the inter-frame difference absolute value is determined to be greater than a motion threshold;
For a pixel for which the interframe difference absolute value is determined to be greater than the luminance change threshold, the luminance gradient value obtained from the differentiation of the interframe luminance of the surrounding pixels of the pixel is the luminance gradient in the comparison determination with the luminance gradient threshold. A fourth step for determining that the pixel is a vertical or oblique edge of a moving subject when the threshold is equal to or greater than a threshold;
A fifth step of counting up the cumulative number of moving pixels by the number of vertical and diagonal edges found on one line;
To return the original progressive video to detect the repeat field by comparing the moving pixel cumulative count number and still pixel cumulative count number of the current field with the moving pixel cumulative count number and still pixel cumulative count number for the past five fields. A sixth step of performing an inverse 23 conversion process of
A film mode detection method comprising:
(Appendix 5)
In Appendix 4,
The film mode detection method, wherein the second step can be replaced with the third step.
(Appendix 6)
In Appendix 4,
The sixth step determines the cumulative number of moving pixels and the cumulative number of still images of the current field and the adjacent fields for the past five fields,
In the determination of the cumulative number of motion pixels, the cumulative number of motion pixels in the first field and the second field are approximated, the cumulative number of motion pixels in the third field and the fourth field are approximated, and the current field and the previous five fields are determined. Both the cumulative number of motion pixels in the field No. 1 indicate the minimum value compared with the cumulative number of motion pixels in the first to fifth fields,
In the determination of the number of still pixels, the number of still pixels accumulated in the current field and the previous five fields is the maximum compared to the number of still pixels accumulated in the first to fifth fields.
A film mode detection method, wherein a current field is detected as a repeat feel based on an AND condition of the result of motion image cumulative number determination and the result of still pixel cumulative determination.
(Appendix 7)
On the computer,
A first step of inputting an input interlaced video signal for a plurality of consecutive fields and calculating an inter-frame difference absolute value;
A second step of comparing the absolute value of the interframe difference with a motion threshold for each pixel of the video signal;
A third step of comparing the inter-frame difference absolute value with a luminance change threshold for a pixel for which the inter-frame difference absolute value is determined to be greater than a motion threshold;
For a pixel for which the interframe difference absolute value is determined to be greater than the luminance change threshold, the luminance gradient value obtained from the differentiation of the interframe luminance of the surrounding pixels of the pixel is the luminance gradient in the comparison determination with the luminance gradient threshold. A fourth step for determining that the pixel is a vertical or oblique edge of a moving subject when the threshold is equal to or greater than a threshold;
A fifth step of counting up the cumulative number of moving pixels by the number of vertical and diagonal edges found on one line;
In order to detect the repeat field by comparing the moving pixel cumulative count number and still pixel cumulative count number of the current field with the moving pixel cumulative count number and still pixel cumulative count number for the past five fields, and to return to the original progressive video A sixth step of performing an inverse 23 conversion process of
A program characterized by having executed.
(Appendix 8)
In Appendix 7,
The program in which the second step can be replaced with the third step.
(Appendix 9)
In Appendix 7,
The sixth step determines the cumulative number of moving pixels and the cumulative number of still images of the current field and the adjacent fields for the past five fields,
In the determination of the cumulative number of motion pixels, the cumulative number of motion pixels in the first field and the second field are approximated, the cumulative number of motion pixels in the third field and the fourth field are approximated, and the current field and the previous five fields are determined. Both the cumulative number of motion pixels in the field No. 1 indicate the minimum value compared with the cumulative number of motion pixels in the first to fifth fields,
In the determination of the still pixel cumulative number, the still pixel cumulative number of the current field and the previous five fields is the maximum value compared with the still pixel cumulative number of the first to fifth fields,
The current field is a program for detecting a repeat field based on an AND condition of the result of the moving image cumulative number determination and the result of the still pixel cumulative determination.

DESCRIPTION OF SYMBOLS 1 1 ₁ , 1 ₂ field memory 2 Frame difference absolute value calculation part 3 Difference absolute value determination part 4 Still pixel accumulation number count part 5 Bias detection part 6 Luminance determination part 7 Edge detection determination part 8 Motion pixel accumulation number count part 9 Approximation Degree determination unit 11 Film material 12 Video image

Claims (4)

An inter-frame difference absolute value calculation unit that inputs the input interlaced video signals for a plurality of continuous fields and calculates an inter-frame difference absolute value;
A difference absolute value determination unit that compares the difference absolute value between frames and a motion threshold value in units of pixels of the video signal;
A luminance determination unit that compares the absolute difference between frames with a luminance change threshold for a pixel for which the absolute difference between frames is determined to be greater than a motion threshold;
A pixel for which the absolute value between the frames is determined to be smaller than a motion threshold value is determined as no motion, and a still pixel cumulative number count unit that increments the still pixel cumulative count number by one;
For a pixel for which the interframe difference absolute value is determined to be greater than the luminance change threshold, the luminance gradient value obtained from the differentiation of the interframe luminance of the surrounding pixels of the pixel is the luminance gradient in the comparison determination with the luminance gradient threshold. An edge detection determination unit that determines that the pixel is a vertical or oblique edge of a moving subject when the pixel is equal to or greater than a threshold;
A moving pixel cumulative number counting unit for counting up the cumulative number of moving pixels by the number of vertical and diagonal edges found on one line;
The moving picture cumulative count number and still pixel cumulative count number of the current field are compared with the moving pixel cumulative count number and still pixel cumulative count number for the past five fields to detect that it is a repeat field, and the original progressive video An approximation determination unit that performs an inverse 23 conversion process to return to
A film mode conversion device comprising: In claim 1,
The degree-of-approximation determination unit determines the cumulative number of motion pixels and the cumulative number of still pixels of adjacent fields for the current field and the past five fields,
In the determination of the cumulative number of motion pixels, the cumulative number of motion pixels in the first field and the second field are approximated, the cumulative number of motion pixels in the third field and the fourth field are approximated, and the current field and the previous five fields are determined. Both the cumulative number of motion pixels in the field No. 1 indicate the minimum value compared with the cumulative number of motion pixels in the first to fifth fields,
In the determination of the still pixel cumulative number, the still pixel cumulative number of the current field and the previous five fields is the maximum value compared with the still pixel cumulative number of the first to fifth fields,
The film mode conversion device according to claim 1, wherein the current field detects a repeat field based on an AND condition of the result of the moving image cumulative number determination and the result of the still pixel cumulative determination. A first step of inputting an input interlaced video signal for a plurality of consecutive fields and calculating an inter-frame difference absolute value;
A second step of comparing the absolute value of the interframe difference with a motion threshold for each pixel of the video signal;
A third step of comparing the inter-frame difference absolute value with a luminance change threshold for a pixel for which the inter-frame difference absolute value is determined to be greater than a motion threshold;
A fourth step of determining that there is no motion for a pixel for which the absolute value between frame differences is smaller than a motion threshold, and incrementing the still pixel cumulative count by one;
For a pixel for which the interframe difference absolute value is determined to be greater than the luminance change threshold, the luminance gradient value obtained from the differentiation of the interframe luminance of the surrounding pixels of the pixel is the luminance gradient in the comparison determination with the luminance gradient threshold. A fifth step for determining that the pixel is a vertical or oblique edge of a moving object when the threshold is equal to or greater than a threshold;
A sixth step of counting up the cumulative number of moving pixels by the number of vertical and diagonal edges found on one line;
The moving picture cumulative count number and still pixel cumulative count number of the current field are compared with the moving pixel cumulative count number and still pixel cumulative count number for the past five fields to detect that it is a repeat field, and the original progressive video A seventh step of performing an inverse 23 conversion process to return to
A film mode conversion method comprising: On the computer,
A first step of inputting an input interlaced video signal for a plurality of consecutive fields and calculating an inter-frame difference absolute value;
A second step of comparing the absolute value of the interframe difference with a motion threshold for each pixel of the video signal;
A third step of comparing the inter-frame difference absolute value with a luminance change threshold for a pixel for which the inter-frame difference absolute value is determined to be greater than a motion threshold;
A fourth step of determining that there is no motion for a pixel for which the absolute value between frame differences is smaller than a motion threshold, and incrementing the still pixel cumulative count by one;
For a pixel for which the inter-frame difference absolute value is determined to be greater than the luminance change threshold, the luminance gradient value obtained from the differentiation of the inter-frame luminance of the surrounding pixels of the pixel is the luminance gradient in the comparison determination with the luminance gradient threshold. A fifth step for determining that the pixel is a vertical or oblique edge of a moving object when the threshold is equal to or greater than a threshold;
A sixth step of counting up the cumulative number of moving pixels by the number of vertical and diagonal edges found on one line;
The moving picture cumulative count number and still pixel cumulative count number of the current field are compared with the moving pixel cumulative count number and still pixel cumulative count number for the past five fields to detect that it is a repeat field, and the original progressive video A seventh step of performing an inverse 23 conversion process to return to
A program characterized by having executed.

Images