Image analysis – Pattern recognition – Classification
Reexamination Certificate
1999-04-21
2001-08-14
Boudreau, Leo (Department: 2621)
Image analysis
Pattern recognition
Classification
C382S115000, C382S118000, C382S133000, C382S165000, C382S238000, C348S384100, C358S461000, C375S240110
Reexamination Certificate
active
06275614
ABSTRACT:
The invention relates generally to a system for processing images and, more particularly, to an apparatus and a method for classifying regions (e.g., blocks or macroblocks) of an image, based on the relative “importance” of the various areas. In turn, the importance information is employed to implement adaptive bit allocation in an encoding environment.
BACKGROUND OF THE INVENTION
An image sequence, such as a video image sequence, typically includes a sequence of image frames or pictures. The reproduction of video containing moving objects typically requires a frame speed of thirty image frames per second, with each frame possibly containing in excess of a megabyte of information. Consequently, transmitting or storing such image sequences requires a large amount of either transmission bandwidth or storage capacity. To reduce the necessary transmission bandwidth or storage capacity, the frame sequence is compressed such that redundant information within the sequence is not stored or transmitted. Television, video conferencing and CD-ROM archiving are examples of applications, which can benefit from efficient video sequence encoding.
Additionally, in an image processing environment where processing resources are limited or constrained by the requirements of a particular application, it is necessary to carefully allocate the available coding bits. For example, in real-time application such as videophone, the talking person's face is typically one of the most important part of an image sequence. Additional “important” regions may include edges or boundaries between distinctive objects
For example, the encoding system in a low bit rate application (e.g., real-time application) must efficiently allocate limited bits to address various demands, i.e., allocating bits to code motion information, allocating bits to code texture information, allocating bits to code shape information, allocating bits to code header information and so on. At times, it may be necessary to allocate available bits such that one parameter will benefit at the expense of another parameter, i.e., spending more bits to provide accurate motion information at the expense of spending less bits to provide texture information. Without information as to which regions in a current frame are particularly important, i.e., deserving of more bits from a limited bit pool, the encoder may not allocate the available bits in the most efficient manner.
Therefore, there is a need in the art for an apparatus and a method for classifying regions of an image, based on the relative “importance” of the various areas and to adaptively use the importance information to implement bit allocation in an encoding environment.
SUMMARY OF THE INVENTION
An embodiment of the present invention is an apparatus and method for classifying regions of an image as important, based on the presence of a human skin-tone, i.e., a skin-tone detector. For example, in videophone applications, the talking person's face is identified as an important region of the frame in an image sequence. The detection of “facial blocks” in a frame is conducted in the chrominance plane. The resulting importance information for the entire image is then collected and represented in an “importance map” or “class map”.
Additionally, other parameters can be used or combined with the skin-tone detector in the generation or refinement of the importance map. Namely, an edge detector for detecting edges of objects in a frame and a motion detector for detecting motion of blocks in a frame can be collectively evaluated to generate the importance map. Furthermore, additional importance information may include temporal importance information, e.g., using the importance map of a previous frame.
Once the importance information is made available, the present invention allocates coding bits in accordance with the importance information. In one embodiment, more bits are allocated to “important” regions as compared to the less “important” regions, so that the important regions are coded with a higher degree of quality. Thus, the relative “importance” of the various areas of a frame is rapidly classified and adaptively used in bit allocation in an encoding environment, e.g., an image encoder.
REFERENCES:
patent: 5309232 (1994-05-01), Hartung et al.
patent: 5412738 (1995-05-01), Brunelli et al.
patent: 5596362 (1997-01-01), Zhou
patent: WO 9602895 (1996-02-01), None
patent: WO 9837811 (1998-09-01), None
Yang, et al. “A Real-Time Face Tracker”, IEEE, May 1996, pp. 142-147.*
Schumeyer R P et al: “Region of Interest priority coding for sign language videoconfrerncing” 1997 IEEE First Workshop on Multimedia Signal Processing (CAT. No. 97TH8256), Proceedings of First Signal Processing Society Workshop on Multimedia Signal Processing, Princeton, NJ, USA, Jun. 23-25, 1997, pp. 531-536, XP002116570 1997, New York, USA, IEEE, USA ISBN: 0-7803-3780-8 p. 533, paragraph 2—p. 534, paragraph 2.
Hance G A et al: “Unsupervised color image segmentation: with application to skin tumor borders” IEEE Engineering in Medicine and Biology Magazine, Jan.-Feb. 1996, IEEE, USA, vol. 15, No. 1, pp. 104-111, XP002116671 ISSN: 0739-5175 p. 104, right-hand column, paragraph 3.
International Search Report corresponding to PCT Application PCT/US99/14508.
“A Highly Efficient System for Automatic Face Region Detection in MPEG Video”, H. Wang and S-F. Chang, IEEE Transactions on Circuits and Systems for Video Technology, vol. 7, No. 4, Aug. 1997, pp. 615-628.
“A Color-Based Classifier for Region Identification in Video”, R. Schumeyer and K. Barner, SPIE Proceedings, voll 3309, Jan. 1999, pp. 189-200.
Krishnamurthy Ravi
Sethuraman Sriram
Boudreau Leo
Burke William J.
Mariam Daniel G.
Sarnoff Corporation
LandOfFree
Method and apparatus for block classification and adaptive... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for block classification and adaptive..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for block classification and adaptive... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2460083