Characterization of motion of moving objects in video

• US 20040223652A1
• Filed: 05/07/2004
• Published: 11/11/2004
• Est. Priority Date: 05/07/2003
• Status: Abandoned Application

• Current Assignee: Ahmet Enis Cetin, Anil Aksay, Behcet Ugur Toreyin, Mehmet Bilgay Akhan
• Original Assignee: Ahmet Enis Cetin, Anil Aksay, Behcet Ugur Toreyin, Mehmet Bilgay Akhan
• Inventors: Akhan, Mehmet Bilgay, Cetin, Ahmet Enis, Toreyin, Behcet Ugur, Aksay, Anil
• Application Number:
  US10/840,817
• Publication Number:
  US 20040223652A1
• Time in Patent Office: Days
• US Class Current: 382/235
• CPC Class Codes: G06T 7/254 involving subtraction of images
• CPC Class Codes: G06T 7/262 using transform domain methods, e.g. Fourier domain methods
• CPC Class Codes: G06V 20/52 Surveillance or monitoring of activities, e.g. for recognising suspicious objects recognising microscopic objects G06V20/69
• CPC Class Codes: G06V 20/54 of traffic, e.g. cars on the road, trains or boats
• CPC Class Codes: G06V 40/20 Movements or behaviour, e.g. gesture recognition recognition of facial expressions G06V40/16

Abstract

A method and system for characterizing the motion of moving objects in video is disclosed. The video can be either compressed or uncompressed form. A typical video scene contains foreground and background objects. Foreground objects temporarily stay in the video. However, a left object or stopped object becomes part of the background scene and remains in the video forever. A plurality of images is inputted to the system in time series. A method and system determines left objects or stopped objects by comparing the background image estimated from the current image of the video with the background estimated from previous images of the video. Difference between the current and previous background images indicates a left or stopped object. Other objects, which do not modify the background image are determined as transitory objects. Background scene of a video can be estimated using the compressed video data as well. If the video is in compressed form, estimating the background in the compressed data domain leads to a computationally efficient method as there is no need to decompress the video. In this case, comparison of the current background scene with the previous background scenes can be carried out in the compressed domain.

Claims

1. A method for characterizing the motion of moving objects and regions in compressed video comprises:

comparing the compressed form of the current image of the video with the estimated compressed form of the background scene, wherein a difference between the compressed form of a current background scene and the compressed form of a past background scene estimated in compressed data domain indicate the existence of at least one stopped object wherein a difference between the current image and a current background image in compressed data domain indicate the existence of at least one object in transition and wherein the nature of the motion or the presence of a left or stopped object in the video is determined without performing data decompression operation.

2. The method of claim 1 wherein the compression of the current image and the background image can be a wavelet, Fourier or Discrete Cosine Transform (DCT) or any other linear transform based method, wherein the nature of the motion or the presence of a left or stopped object in the video is determined without performing inverse transformation operation.

3. The method of claim 1 wherein the data compression method comprises a block based method including DCT and vector quantisation based methods, and methods performing transformations in blocks of data forming image frames, wherein the nature of the motion or the presence of a left or stopped object in the video is determined without performing decompression operation.

4. The method of claim 1, wherein said comparing step comprises matching the predetermined area in the wavelet transform of one image with the predetermined area in the wavelet transform of the next image by shifting as one unit in the wavelet domain, calculating the difference of wavelet coefficient values between the predetermined area in the wavelet transform of the one image and each matched area of the wavelet transform of the next image, and calculating an evaluation value of the difference of the wavelet coefficient value.

5. The method of claim 1 wherein the threshold values determining the moving wavelet coefficients are estimated in a recursive manner from the threshold value used in previous comparison, and difference of the previous value of the wavelet coefficient and estimated wavelet coefficient of the background in wavelet compressed video, wherein the system updates the threshold values by itself without requiring any predefined threshold values except an initial threshold value.

6. The method of claim 1 wherein the locations of moving objects on the original image data domain are estimated by determining the indices of the image pixels producing the wavelet coefficients of the current image frame differing from the wavelet coefficients of the estimated background in wavelet compressed video.

7. The method of claim 1 wherein the locations of left or stopped objects on the original image data domain are estimated by determining the indices of the background image pixels producing the wavelet coefficients of the current background image differing from the wavelet coefficients of a past background image in wavelet compressed video.

8. The method of claim 1 wherein block based compression schemes employ are of a DCT, Discrete Fourier or any other linear transform, and a coarse form of the background image can be estimated from the DC coefficients of transformed image blocks.

9. The method of claim 8 wherein a background estimation scheme employing DC coefficients of transformed image blocks is utilized.

10. The method of claim 1 wherein a block based video coding scheme is utilized, and a compressed form of the background image can be estimated from blocks which do not move or equivalently from blocks whose motion vectors are below a threshold.

11. The method of claim 10, a compressed form of the background image estimated from compressed form of image blocks, which do not move over time is utilized.

12. The method of claim 11 wherein a video compression schemes employs a DCT, Discrete Fourier or any other linear transform, and a compressed form of the background image can be estimated by averaging the transform domain data over time.

13. The method of claim 1 wherein a background estimation scheme by averaging the transform domain data over time is utilized.

14. A computer readable medium containing program instructions for characterizing the motion of moving objects and regions in compressed video, the program instructions for,

comparing the compressed form of the current image of the video with the estimated compressed form of the background scene, wherein a difference between the compressed form of a current background scene and the compressed form of a past background scene estimated in compressed data domain indicate the existence of at least one stopped object wherein a difference between the current image and a current background image in compressed data domain indicate the existence of at least one object in transition and wherein the nature of the motion or the presence of a left or stopped object in the video is determined without performing data decompression operation.

15. The computer readable medium of claim 14 wherein the compression of the current image and the background image can be a wavelet, Fourier or Discrete Cosine Transform (DCT) or any other linear transform based method, wherein the nature of the motion or the presence of a left or stopped object in the video is determined without performing inverse transformation operation.

16. The computer readable medium of claim 14 wherein the data compression method comprises a block based method including DCT and vector quantisation based methods, and methods performing transformations in blocks of data forming image frames, wherein the nature of the motion or the presence of a left or stopped object in the video is determined without performing decompression operation.

17. The computer readable medium of claim 14 wherein said comparing step comprises matching the predetermined area in the wavelet transform of one image with the predetermined area in the wavelet transform of the next image by shifting as one unit in the wavelet domain, calculating the difference of wavelet coefficient values between the predetermined area in the wavelet transform of the one image and each matched area of the wavelet transform of the next image, and calculating an evaluation value of the difference of the wavelet coefficient value.

18. The computer readable medium of claim 14 wherein the threshold values determining the moving wavelet coefficients are estimated in a recursive manner from the threshold value used in previous comparison, and difference of the previous value of the wavelet coefficient and estimated wavelet coefficient of the background in wavelet compressed video, wherein the system updates the threshold values by itself without requiring any predefined threshold values except an initial threshold value.

19. The computer readable medium of claim 14 wherein the locations of moving objects on the original image data domain are estimated by determining the indices of the image pixels producing the wavelet coefficients of the current image frame differing from the wavelet coefficients of the estimated background in wavelet compressed video.

20. The computer readable medium of claim 14 wherein the locations of left or stopped objects on the original image data domain are estimated by determining the indices of the background image pixels producing the wavelet coefficients of the current background image differing from the wavelet coefficients of a past background image in wavelet compressed video.

21. The computer readable medium of claim 14 wherein block based compression schemes employ are of a DCT, Discrete Fourier or any other linear transform, and a coarse form of the background image can be estimated from the DC coefficients of transformed image blocks.

22. The computer readable medium of claim 21 wherein a background estimation scheme employing DC coefficients of transformed image blocks is utilized.

23. The computer readable medium of claim 14 wherein a block based video coding scheme is utilized, and a compressed form of the background image can be estimated from blocks which do not move or equivalently from blocks whose motion vectors are below a threshold.

24. The computer readable medium of claim 23, a compressed form of the background image estimated from compressed form of image blocks, which do not move over time is utilized.

25. The computer readable medium of claim 24 a video compression scheme employs a DCT, Discrete Fourier or any other linear transform, and a compressed form of the background image can be estimated by averaging the transform domain data over time.

26. The computer readable medium of claim 22 wherein a background estimation scheme by averaging the transform domain data over time is utilized.