Video image feature generation system and video image feature generation method

1. A video image feature generation system comprising:

a processor; and

a memory which stores a plurality of instructions, which when executed by the processor, cause the processor to execute,

extracting a frame feature value featuring a frame, which is a unit of an input video image, based on a pixel value of the frame; and

generating a phase of each frequency as a video image feature based on at least two frequencies, the frame feature value obtained in the extracting, and generation information for generating phases of the frequencies according to the frequencies and the frame feature value.

2. The system according to claim 1 , further comprising:

selecting the at least two frequencies from a result of a Fourier transform performed on the frame feature value obtained in the extracting.

3. The system according to claim 2 ,

wherein the extracting includes extracting a time series of frame feature values based on pixel values on a plurality of frames included in the video image,

wherein the selecting includes selecting at least two frequencies from a frequency spectrum based on the time series of frame feature values obtained in the extracting,

wherein the generating includes generating a time series of phases of each frequency as a video image feature based on the at least two frequencies obtained in the selecting, the time series of frame feature values obtained in the extracting, and the generation information.

4. The system according to claim 1 ,

wherein the extracting includes extracting the frame feature value based on all pixel values included in the frame.

5. The system according to claim 4 ,

wherein the extracting includes extracting the frame feature value based on a mean value of all pixel values included in the frame.

6. The system according to claim 4 ,

wherein the extracting includes extracting the frame feature value based on a variance of all pixel values included in the frame.

7. The system according to claim 2 ,

wherein the extracting includes extracting a pixel value as the frame feature value at each location on at least one frame included in the video image,

wherein the selecting includes selecting at least two frequencies including a horizontal component and a vertical component from a frequency domain based on the frame feature value obtained by the extracting,

wherein the generating includes generating a phase of each frequency as a video image feature based on the at least two frequencies obtained by the selecting, the frame feature value obtained by the extracting, and the generation information.

8. A video image feature generation method comprising:

extracting a frame feature value featuring a frame, which is a unit of an input video image, based on a pixel value of the frame; and

generating, by a computer processor, a phase of each frequency as a video image feature based on at least two frequencies, the frame feature value obtained in the extracting, and generation information for generating phases of the frequencies according to the frequencies and the frame feature value.

9. The method according to claim 8 , further comprising:

selecting the at least two frequencies from a result of a Fourier transform performed on the frame feature value obtained in the extracting.

10. The method according to claim 9 ,

wherein the extracting includes extracting a time series of frame feature values based on pixel values on a plurality of frames included in the video image,

wherein the selecting includes selecting at least two frequencies from a frequency spectrum based on the time series of frame feature values obtained in the extracting,

wherein the generating includes generating a time series of phases of each frequency as a video image feature based on the at least two frequencies obtained in the selecting, the time series of frame feature values obtained in the extracting, and the generation information.

11. The method according to claim 8 ,

wherein the extracting includes extracting the frame feature value based on all pixel values included in the frame.

12. The method according to claim 11 ,

wherein the extracting includes extracting the frame feature value based on a mean value of all pixel values included in the frame.

13. The method according to claim 11 ,

wherein the extracting includes extracting the frame feature value based on a variance of all pixel values included in the frame.

14. The method according to claim 9 ,

wherein the extracting includes extracting a pixel value as the frame feature value at each location on at least one frame included in the video image,

wherein the selecting includes selecting at least two frequencies including a horizontal component and a vertical component from a frequency domain based on the frame feature value obtained by the extracting,

wherein the generating includes generating a phase of each frequency as a video image feature based on the at least two frequencies obtained by the selecting, the frame feature value obtained by the extracting, and the generation information.

15. A non-transitory computer-readable storage medium storing a video image feature generation program that causes a computer to execute a process comprising:

extracting a frame feature value featuring a frame, which is a unit of an input video image, based on a pixel value of the frame; and

generating a phase of each frequency as a video image feature based on at least two frequencies, the frame feature value obtained in the extracting, and generation information for generating phases of the frequencies according to the frequencies and the frame feature value.

16. A video image comparison device comprising:

a processor; and

a memory which stores a plurality of instructions, which when executed by the processor, cause the processor to execute,

extracting a frame feature value featuring a frame for each of first and second video images based on a pixel value of each frame, each video image including frames as units forming the image;

generating a phase of each frequency as a first video image feature for the first video image and a phase of each frequency as a second video image feature for the second video image based on at least two frequencies, the frame feature value extracted by the extraction unit, and generation information for generating phases of the frequencies according to the frequencies and the frame feature value; and

making a comparison between the first and second video image features generated by the generation unit.