Motion detection method and apparatus

• US 5,937,097 A
• Filed: 12/09/1996
• Issued: 08/10/1999
• Est. Priority Date: 12/21/1995
• Status: Expired due to Fees

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Lennon, Alison
• Primary Examiner(s): Couso, Jose L.
• Assistant Examiner(s): Do, Anh Hong
• Application Number:
  US08/762,254
• Time in Patent Office: 974 Days
• Field of Search: 382/236, 348/407, 348/409, 358/261.2, 358/261.1
• US Class Current: 382/236
• CPC Class Codes: G06T 7/262 using transform domain methods, e.g. Fourier domain methods
• CPC Class Codes: G09G 2310/04 Partial updating of the display screen
• CPC Class Codes: G09G 2320/0261 in the context of movement of objects on the screen or movement of the observer relative to the screen
• CPC Class Codes: G09G 3/3629 using liquid crystals having memory effects, e.g. ferroelectric liquid crystals

Abstract

The present invention provides a method and apparatus for detecting motion in an input sequence of images. Each image comprises a number of lines and each line has a number of pixels. Values of a region of pixels of a current image are transformed to determine a set of representation values. The set of representation values are indicative of the region of pixels at a number of resolution levels. The set of representation values of the region of pixels of the current image are compared with a set of representation values of a corresponding region of pixels from a stored previous image. Based on the comparing step, a difference measure is determined between the region of pixels and the corresponding region of pixels that is indicative of the presence or absence of motion in the current image.

Claims

1. An apparatus for detecting motion in an input sequence of images, each image comprising a plurality of lines and each line comprising a plurality of pixels, said apparatus comprising:

coding means for transforming values of pixels within at least one region of a current image to form a set of representation values at a plurality of different resolution levels;

means for comparing and iteratively determining a difference measure between said representation values for at least one of said plurality of different resolution levels and corresponding representation values from a stored set of representation values of a preceding image of said input sequence; and

difference means for determining a difference value from difference measures determined by said means for comparing, said difference value being generated by adding together difference measures for said plurality of resolution levels and indicating motion between said current image and said preceding image.

2. The apparatus as claimed in claim 1, wherein said transforming means comprises means for wavelet transforming said values of said pixels to form said set of representation values.

3. The apparatus as claimed in claim 2, wherein said wavelet transforming means is adapted to implement one of the group consisting of Haar and Daubechies transforms.

4. The apparatus as claimed in claim 1 or 2, wherein the transform is repeatedly applied to said values of said pixels within said at least one region to form the set of representation values at each resolution level of said plurality of resolution levels.

5. The apparatus as claimed in claim 1, wherein said set of representation values are stored in a frame buffer.

6. The apparatus as claimed in claim 1, wherein said values of pixels are color values of the pixels.

7. The apparatus as claimed in claim 1, wherein said values of pixels are luminance and chrominance values of pixels.

8. The apparatus as claimed in claim 1, wherein said means for comparing forwards said difference measure to an absolute value determination means which forms an absolute difference, said absolute difference being thresholded and forwarded to a first summing means to form a level difference measure.

9. The apparatus as claimed in claim 8, wherein a second summing means sums, for all resolution levels of said region of pixels, a product of a weight value and said level difference measure.

10. A method for detecting motion in an input sequence of images, each image comprising a plurality of lines and each line comprising a plurality of pixels, said method comprising the steps of:

transforming values of a region of pixels to determine a set of representation values, said set of representation values being indicative of the region of pixels at a plurality of resolution levels;

comparing and iteratively determining a difference measure between the set of representation values of the region of pixels, of a current image, and a set of representation values of a corresponding region of pixels from a preceding image; and

determining a difference value from difference measures obtained in said comparing step, said difference value being generated by adding together difference measures for said plurality of resolution levels and indicating motion between said current image and said preceding image.

11. The method as claimed in claim 10, wherein said transforming step is repeatedly applied to form a set of representation values for each resolution level of said region of pixels.

12. The method as claimed in claim 11 wherein said comparing step further includes the step of comparing said set of representation values for each level with a corresponding set of representation values at the respective resolution level of the corresponding region of pixels.

13. The method as claimed in claim 10 wherein the transform is a wavelet transform.

14. A method as claimed in claim 10, wherein said difference measure is forwarded to an absolute value determination means which forms an absolute difference, said absolute difference being thresholded and forwarded to a first summing means to form a level difference measure.

15. A method as claimed in claim 14, wherein a product of a weight value and said difference value, for all resolution levels of said region of pixels, is summed.

16. An apparatus for providing a display signal from an input sequence of images, said apparatus comprising:

memory for storing image data;

means for generating said display signal, wherein said generating means partially updates a display means for each region of a current image; and

means for detecting motion in a current image of said input sequence, said detecting means being coupled to said memory and said generating means, said detecting means processing said current image as a plurality of blocks, said detecting means comprising;

means for decomposing a block of pixel values of said current image to provide a set of representation coefficients at a plurality of different resolution levels;

means for iteratively computing a difference for each resolution level between said set of representation coefficients of said current image and corresponding representation coefficients of a previous image stored in said memory; and

means for measuring a block difference for said current block dependent on a plurality of differences, wherein said block difference measure indicates motion in said current block;

wherein said generating means generates said display signal using said image data stored in said memory in response to said block difference measure, wherein said display signal contains a block of pixel values dependent upon said current block of pixel values.

17. The apparatus according to claim 16, wherein said decomposing means comprises means for wavelet transforming said block of pixel values to provide said set of representation coefficients.

18. The apparatus according to claim 17, wherein the number of resolution levels of said set of representation coefficients corresponds to the number of repeated applications of said wavelet transformation to said block of pixel values.

19. The apparatus according to claim 18, wherein said wavelet transformation is one of the group consisting of Haar transform and Daubechie transform.

20. The apparatus according to claim 16, wherein said difference computing means comprises:

means for determining the absolute value of said difference for each resolution level.

21. The apparatus according to claim 16, wherein said measuring means comprises:

means for weighting each of said differences; and

means for summing said weighted differences.

22. The apparatus according to claim 16, wherein said input sequence of images is a video signal having a first data rate, and said display signal has a second data rate, wherein said second data rate is lower than said first data rate.

23. The apparatus according to claim 22, wherein said display signal is adapted for display on a liquid crystal display.

24. The apparatus according to claim 23, wherein said liquid crystal display is a ferro-electric liquid crystal display.

25. An apparatus as claimed in claim 16, wherein said difference is output to an absolute value determination means and wherein said absolute value determination means outputs an absolute difference to a first summer means to form a level difference measure.

26. An apparatus as claimed in claim 25, wherein a second summer means sums, for all resolution levels of said region of pixels, a product of a weight value and said level difference measure.

27. A method for providing a display signal from an input sequence of images, said method comprising the steps of:

detecting motion in a current image of said input sequence, wherein said current image is processed as a plurality of blocks, said detecting step further comprising the steps of;

decomposing a block of pixel values of said current image to provide a set of representation coefficients at a plurality of different resolution levels;

iteratively determining a difference measure for each resolution level between said set of representation coefficients of said current image and corresponding representation coefficients of a previous image stored in a memory; and

measuring a block difference for said current block dependent on a plurality of differences, wherein said block difference measure indicates motion in said current block; and

generating said display signal so as to partially update a display means using said image data stored in said memory in response to said block difference measure, wherein said display signal comprises a block of pixel values dependent upon said current block of pixel values.

28. The method according to claim 27, wherein said decomposing step comprises the step of wavelet transforming said block of pixels values to provide said set of representation coefficients.

29. The method according to claim 28, wherein the number of resolution levels of said set of representation coefficients corresponds to the number of repeated applications of said wavelet transformation to said block of pixel values.

30. The method according to claim 29, wherein said wavelet transformation is one of the group consisting of Haar transform and Daubechie transform.

31. The method according to claim 27, wherein said difference computing step comprises the further step of:

determining the absolute value of said difference for each resolution level.

32. The method according to claim 27, wherein said measuring step comprises the further steps of:

weighting each of said differences; and

summing said weighted differences.

33. The method according to claim 27, wherein said input sequence of images is a video signal having a first data rate, and said display signal has a second data rate, wherein said second data rate is lower than said first data rate.

34. The method according to claim 33, wherein said display signal is adapted for display on a liquid crystal display.

35. The method according to claim 34, wherein said liquid crystal display is a ferro-electric liquid crystal display.

36. A method as claimed in claim 27, wherein said difference is forwarded to an absolute value determination means which forms an absolute difference, said absolute difference being thresholded and forwarded to a first summing means to form a level difference measure.

37. A method as claimed in claim 36, wherein a second summing means sums, for all resolution levels of said region of pixels, a product of a weight value and said level difference means.

38. A computer software system having a set of instructions for execution on processing means to detect motion between frames of an input video signal, said set of instructions formed into a plurality of process modules comprising:

a transformation process for transforming a block of pixels of a current image frame to form a set of representation values at a plurality of different resolution levels;

a comparison process for comparing and iteratively determining a difference measure between said set of representation values of the current image frame and a set of representation values of a corresponding block of pixels of a previous image frame; and

a difference process for measuring a block difference between said block of said current image frame and said corresponding block of said previous image frame, utilizing said difference measure, said process including adding together difference measures for said plurality of resolution levels, wherein said block difference indicates motion in said current image frame.

39. The computer software system according to claim 38, wherein said transforming process repeatedly applies a wavelet transformation, wherein the number of repeated applications corresponds to the number of different resolutions levels.

40. The computer software system according to claim 39, wherein said wavelet transformation is one of the group consisting of Haar and Daubechies transforms.

41. A computer software system as claimed in claim 38, wherein a result of said comparison process is forwarded to an absolute value determination means which forms an absolute difference, said absolute difference being thresholded and forwarded to a first summing means to form a level difference measure.

42. A computer software system as claimed in claim 41, wherein a second summing means sums, for all resolution levels of said region of pixels, a product of a weight value and said level difference measure.