KR100982509B1 - Film mode detection method and apparatus - Google Patents

Abstract

The present invention relates to video signal processing, and more particularly, to a method and apparatus for detecting a film mode of an input video signal. The film mode detection method of the present invention comprises the steps of: (a) receiving motion prediction mode information of each field of the interlaced video signal; (b) counting the number of macro blocks having a motion vector of (0, 0) by performing motion prediction on the respective fields under the motion prediction mode; (c) comparing the counted value with a predetermined threshold value and outputting the value as 0 or 1; And (d) determining the film mode by looking for the periodicity of the output values. Using the method of the present invention, since the additional information generated by the moving picture encoder is used, the hardware is not complicated and no additional calculation process is required, which is efficient.

Description

Film mode detection method and apparatus

1 is a diagram illustrating the conversion of 24 movie frames per second to 60 fields per second using the 3: 2 pulldown method.

2 is a diagram illustrating conversion of an interlaced field into a progressive frame.

3 is a diagram illustrating detecting a film mode from an input image having an IPPP structure.

4 is a diagram illustrating detecting a film mode from an input image having an IBP structure.

5 is a flowchart of the film mode detection method described above.

6 is a block diagram of a film mode detection apparatus.

The present invention relates to video signal processing, and more particularly, to a method and apparatus for detecting a film mode of an input video signal.

Images played in movies, TV, DVD, etc. have a different frame structure. When the input image source is a movie, it is called a film mode. A movie is composed of 24 progressive frames per second, and a TV image or a DVD video is composed of 60 interlaced fields per second. Therefore, in order to broadcast a movie produced on a movie, it is necessary to convert 24 progressive frames per second into 60 interlaced fields per second. Performing this conversion is called telecine, and a method called 3: 2 pull down is used in converting a movie into an NTSC TV image.

1 is a diagram illustrating the conversion of 24 movie frames per second to 60 fields per second using the 3: 2 pulldown method.

Referring to FIG. 1, the top field 111 and the bottom field 112 are extracted from one frame 110, and the top field 113 is extracted again. Then, the bottom field 121 and the top field 122 are extracted in the next frame 120. Through this process, two progressive frames are converted into five interlaced fields. Alternatively, after extracting the bottom field 131, the top field 132, and the bottom field 133 in one frame 130, the top field and the bottom field are extracted from the next frame. Similarly, this process converts two progressive frames into five interlaced fields.

When the movie screen is converted into 60 interlaced fields by the above-described 3: 2 pull-down method, 60 progressive frames may be combined by combining the top field and the bottom field.                         

2 is a diagram illustrating conversion of an interlaced field into a progressive frame.

Given an image converted from the movie screen by the 3: 2 pulldown method, a progressive frame before the 3: 2 pulldown is performed by combining the top field and the bottom field. If you play it like this, you can get better image quality and you can get smooth motion video even when playing in slow motion. In addition, by storing a movie having 24 Frame / second at 60 Field / second, it is possible to prevent waste of storage space. However, it is possible to perform different processing according to the type of input image only by determining whether the input image is a 3: 2 pull-down image or an original image composed of an interlaced field. Detection of these video modes is also required for PAL or other video signals other than NTSC.

There are two conventional methods for detecting the film mode, depending on the detection characteristics used. The first is to use a pattern of sum of absolute difference (SAD) between fields calculated every two field (one frame) time. More specifically, there are a method using the characteristic of the period of this SAD is 5 and a method of obtaining the difference signal from the previous SAD and using the characteristic of the period of this difference signal is 5 to obtain a more periodic SAD pattern.

The second method is to use a pattern characteristic of interfield movement between field times. More specifically, the method using the sum of the motion vector of the block and the sum of the previously obtained sum should be within a certain range, and the magnitude of the motion of the entire field by calculating and adding the magnitude of the motion of the pixel. There is a method of detecting the film mode by comparing with a binary sequence (for example, 100010) when obtaining a signal and converting the signal into a digital pattern.

The method using the SAD pattern uses a difference between frame information of the same field, and the method using a motion vector uses a motion vector between immediately adjacent fields. However, there are still difficulties in finding the correct pattern.

An object of the present invention is to provide a method and apparatus for more accurately detecting a film mode by using difference information and motion vector information between fields of different types as well as fields of the same type.

In order to achieve the above object, the film mode detection method according to the present invention comprises the steps of: (a) receiving motion prediction mode information of each field of an interlaced video signal; (b) counting the number of macro blocks having a motion vector of (0, 0) by performing motion prediction on the respective fields under the motion prediction mode; (c) comparing the counted value with a predetermined threshold value and outputting the value as 0 or 1; And (d) determining the film mode by looking for the periodicity of the output values.

In order to achieve the above object, the film mode detecting apparatus according to the present invention includes a ME information input unit for receiving motion prediction mode information of each field of an interlaced video signal; A coefficient unit for performing motion prediction on each of the fields under the motion prediction mode and counting the number of macro blocks having a motion vector of (0, 0); A comparison unit comparing the counted value with a predetermined threshold value and outputting the value as 0 or 1; And a pattern detection unit that determines a film mode by looking for periodicity of the output value.

In order to achieve the above object, the present invention provides a computer-readable recording medium recording a program for executing the method on a computer.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

The film mode detection method of the present invention is used in a detelecine method operating in an encoder, and operates under the following basic assumptions. That is, the input image is an image in which the top field comes out first. In addition, the motion estimation module (ME) of the encoder performs the ME process as if it is a P frame for the I-frame, and assumes that the actual encoding is an I-frame. This is because motion information is also required for the I-frame. In an encoder, the ME module is usually always running, so this assumption is easy. Meanwhile, the present invention works well in the case of GOP (Group of Picture) images of the IPPP structure and the IBP structure, unless a special ME module is used.

3 is a diagram illustrating detecting a film mode from an input image having an IPPP structure.

In the case of an image of an interlaced field, both field ME and frame ME are performed, and a motion prediction (ME) of a less distortion type is selected. In the P frame composed of F ^T _{n + 1} and F ^B _{n + 2} , P3 310, the field ME is T2T 311 which is a motion prediction (ME) from the top field to the top field, as can be seen with reference to FIG. 3. T2B 312, which is the motion prediction from the top field to the bottom field, B2T 313, which is the motion prediction from the bottom field to the top field, and B2B 314, which is the motion prediction from the bottom field to the bottom field. The top field selects a mode with less distortion among the T2T 311 and the T2B 312, and the bottom field selects a mode with less distortion among the B2T 313 and the B2B 314.

Referring to FIG. 3, most of the macro blocks are selected in the T2T mode in the motion prediction of the top field of P3 310. At that time, the motion vector (MV) is likely to be (0, 0). This is because the top fields are all equal to F ^T _{n + 1} in the T2T 311 motion prediction of P3 310. In the top field motion prediction of P3 310, let N ^T ₃ be the number of macroblocks in T2T mode and the motion vector (MV) is (0, 0), ideally N ^T ₃ is the value of the macroblock in the field. Same as number. The value of N ^T _{3 at} that time becomes the maximum value. Although not ideal, it will be close to the maximum.

In motion estimation of the bottom field, the number of macroblocks having a motion vector of (0, 0) in B2B mode is calculated. And let's call N ^B ₃ the number of macro blocks at that time. In this way, if the field-wise motion prediction is performed for P4 and P5 frames, successive count values for each field N ^T ₃ , N ^B ₃ , N ^T ₄ , N ^B ₄ , N ^T ₅ , N ^B You can get ₅ As can be seen from FIG. 3, since the fields are identical to each other in the T2T 311 prediction mode in the P3 field 310 and the fields are identical to each other in the B2B 321 prediction mode in the P5 field 320, N ^T _3. And N ^B ₅ have a larger value than the other values. Therefore, if an appropriate threshold value is set, and if it is larger than the threshold value, it is 1, and if it is not large, it is 0, and the comparison result is a pattern such as "0, 1, 0, 0, 0, 0, 1". If the input image is a film mode, a value of 1 appears to be 5 periods. Therefore, by calculating the period value, it is possible to know whether the input image is in the film mode.

4 is a diagram illustrating detecting a film mode from an input image having an IBP structure.

Only the field ME in the B-frame can obtain a periodic pattern as described above. The forward motion prediction of the top field in the B2 frame 410 consists of T2T_F 411 and T2B_F 412. The forward motion prediction of the bottom field in the B2 frame 410 is composed of the B2T_F 413 and the B2B_F 414. The backward motion prediction of the top field is composed of T2T_B 415 and T2B_B 416. The backward motion prediction of the bottom field is composed of B2T_B 417 and B2B_B 418.

In B2 frame 410, the number of macroblocks in T2T mode and motion vector (0, 0) in the top field forward motion prediction is N ₂ ^{T (F)} , and in T2T mode and motion vector in backward motion prediction Let 0, 0) be the number of macroblocks N ₂ ^{T (B)} . The number of macroblocks in B2B mode and motion vectors of (0, 0) in the forward field prediction of the bottom field is called N ₂ ^{B (F)} . Let N ₂ ^{B (F)} be the number of macroblocks. Referring to Figure 4 it can be seen that the value of N ₂ ^{T (B)} in the film mode is very large. Also for the B4 frame 420, the number N ₂ ^{T (F)} , N ₄ ^{T (B)} , N ₄ ^{B (F), and} N ₄ ^{B (F) described above} are calculated.

List them in order: N ₂ ^{T (F)} , N ₂ ^{B (F)} , N ₂ ^{T (B)} , N ₂ ^{B (F} , N ₂ ^{T (F)} , N ₄ ^{B (F)} , N ₄ ^{T (B)} , N ₄ ^{B (F)} and each of these values is listed in comparison with the appropriate threshold value, and becomes "0, 0, 1, 0, 0, 0, 0, 1". In this case, the value of 1 indicates that the period is 5. Therefore, calculating the period value indicates whether the input image is in film mode.

While the encoder is operating in film mode, it continuously checks whether the input image is in film mode. If the above pattern is broken, it is determined that the image is not in the film mode and the mode is switched to the interlaced image encoding mode. In the film mode, as described with reference to FIG. 2, the encoding is performed after converting to the progressive format through an interlaced-to-progressive conversion process. Motion prediction in the encoder is performed both field ME and frame ME. Therefore, there are field ME results between fields of the same type, whether IPPP or IBP. Therefore, based on the result, it is possible to accurately derive the motion prediction result of the progressive frame. Eventually the motion prediction is done as if it is an interlaced frame, while the encoding is performed on progressive frames.

The above method is not applicable only to the encoder. That is, when an input having an IPPP structure comes in, it can operate at the front end of the encoder or the rear end of the decoder. Since the motion prediction process may be a computational burden, the amount of computation is determined by checking only the motion vector is (0,0) and counting the number of macroblocks in case of T2T or the number of macroblocks in case of B2B. Can be reduced.

5 is a flowchart of the film mode detection method described above.

The motion prediction mode information of each field of the interlaced video signal is input (S510). Since motion prediction is continuously being performed, motion prediction information in each field may be input. As described above with reference to FIG. 3, when the interlaced video signal has an IPPP structure, motion prediction mode information of the video signal for each P frame is received for each field, and as described above with reference to FIG. 4. When the interlaced video signal has an IBP structure, motion prediction mode information of the video signal for each B frame is received for each field.

Then, motion prediction is performed on each field under the motion prediction mode, and the number of macroblocks having a motion vector of (0, 0) is counted (S520). Next, the counted value is compared with the predetermined threshold value and output as 0 or 1 (S530). In operation S540, the film mode is determined by looking for the periodicity of the output value. As described above, when the period is 5, it may be determined as the film mode.

6 is a block diagram of a film mode detection apparatus.

The film mode detecting apparatus includes an ME information input unit 610, a counter 620, a comparator 630, and a pattern detector 640.

The ME information input unit 610 receives motion prediction mode information of each field of the interlaced video signal. Since motion prediction is continuously being performed, motion prediction information in each field may be input. As described above with reference to FIG. 3, when the interlaced video signal has an IPPP structure, motion prediction mode information of the video signal for each P frame is received for each field, and as described above with reference to FIG. 4. When the interlaced video signal has an IBP structure, motion prediction mode information of the video signal for each B frame is received for each field.

The coefficient unit 620 performs motion prediction on the respective fields under the motion prediction mode, and counts the number of macro blocks having a motion vector of (0, 0). The comparator 630 outputs 0 or 1 by comparing the counted value with a predetermined threshold value. The pattern detector 640 determines the film mode by searching the periodicity of the output values.

The invention can also be embodied as computer readable code on a computer readable recording medium. Computer-readable recording media include all types of recording devices that store data that can be read by a computer system. Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage, and the like, and may also be implemented in the form of a carrier wave (for example, transmission over the Internet). Include. The computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion.

So far I looked at the center of the preferred embodiment for the present invention. Those skilled in the art will appreciate that the present invention can be implemented in a modified form without departing from the essential features of the present invention. Therefore, the disclosed embodiments should be considered in descriptive sense only and not for purposes of limitation. The scope of the present invention is shown in the claims rather than the foregoing description, and all differences within the scope will be construed as being included in the present invention.

As described above, the present invention uses the additional information generated by the video encoder, so that the hardware is not complicated and a separate calculation process is not required.

Claims (11)

(a) receiving motion prediction mode information of each field of the interlaced video signal; (b) counting the number of macro blocks having a motion vector of (0, 0) by performing motion prediction on the respective fields under the motion prediction mode; (c) comparing the counted number of macroblocks obtained in step (b) with a predetermined threshold and outputting 0 or 1; And (d) If the period in which the value of counting the number of macroblocks when the fields are identical to each other in the motion prediction mode based on the value output in step (c) has a predetermined period, And judging as a film mode, The threshold value is 0 or 1 such that the values obtained by counting the number of the macroblocks obtained in step (b) are patterns representing the values of the number of the macroblocks when the fields are the same in the motion prediction mode. Film mode detection method, characterized in that the set value to output. The method of claim 1, wherein step (a) When the interlaced video signal has an IPPP structure, motion mode prediction information of the video signal for each P frame is input for each field. The method of claim 1, wherein step (a) And the motion prediction mode information of the video signal for each B frame is input for each field when the interlaced video signal has an IBP structure. The method according to claim 2 or 3, And the interlaced video signal is an interlaced video signal having a top field first. The method of claim 1, In the step (d), when the period in which 1 appears in the value output in the step (c) is 5, the input image signal is determined as a film mode. A motion prediction mode (ME) information input unit for receiving motion prediction mode information of each field of the interlaced video signal; A coefficient unit for performing motion prediction on each of the fields under the motion prediction mode and counting the number of macro blocks having a motion vector of (0, 0); A comparator for comparing the counted number of macroblocks with a predetermined threshold and outputting 0 or 1; And If the period in which the value of counting the number of macroblocks when the fields are identical to each other in the motion prediction mode has a preset period based on the value output from the comparator, determining the input image signal as the film mode. Including a pattern detection unit, The threshold value is 0 or 1 so that the values obtained by counting the number of the macroblocks obtained by the coefficient unit are patterns representing the values of the number of the macroblocks when the fields are the same in the motion prediction mode. Film mode detection device, characterized in that the set value to. The method of claim 6, wherein the motion prediction mode (ME) information input unit When the interlaced video signal has an IPPP structure, the film mode detection device according to claim 1, wherein the motion prediction mode information of the video signal for each P frame is input for each field. The method of claim 6, wherein the motion prediction mode (ME) information input unit When the interlaced video signal has an IBP structure, the film mode detection device according to claim 1, wherein the motion prediction mode information of the video signal for each B frame is input for each field. 9. The method according to claim 7 or 8, And the interlaced video signal is an interlaced video signal having a top field first. The method of claim 6, And the pattern detector determines the input image signal as a film mode when a period in which 1 appears in a value output from the comparator is five. A computer-readable recording medium having recorded thereon a program for executing the method according to any one of claims 1 to 3 and 5.

Images