Image analysis – Image segmentation – Distinguishing text from other regions
Reexamination Certificate
1999-08-09
2003-08-19
Do, Anh Hong (Department: 2721)
Image analysis
Image segmentation
Distinguishing text from other regions
C382S173000, C382S190000
Reexamination Certificate
active
06608930
ABSTRACT:
TECHNICAL FIELD OF THE INVENTION
The present invention is directed, in general, to video processing systems and, more specifically, to a system for analyzing and characterizing a video stream based on the attributes of text detected in the content of the video.
BACKGROUND OF THE INVENTION
The advent of digital television (DTV), the increasing popularity of the Internet, and the introduction of consumer multimedia electronics, such as compact disc (CD) and digital video disc. (DVD) players, have made tremendous amounts of multimedia information available to consumers. As video content becomes readily available and products for accessing it reach the consumer market, searching, indexing and identifying large volumes of multimedia data becomes even more challenging and important.
Systems and methods for indexing and classifying video have been described in numerous publications, including: M. Abdel-Mottaleb et al., “CONIVAS: Content-based Image and Video Access System,” Proceedings of ACM Multimedia, pp. 427-428, Boston (1996); S-F. Chang et al., “VideoQ: An Automated Content Based Video Search System Using Visual Cues,” Proceedings of ACM Multimedia, pp. 313-324, Seattle (1994); M. Christel et al., “Informedia Digital Video Library,” Comm. of the ACM, Vol. 38, No. 4, pp. 57-58 (1995); N. Dimitrova et al., “Video Content Management in Consumer Devices,” IEEE Transactions on Knowledge and Data Engineering (Nov. 1998); U. Gargi et al., “Indexing Text Events in Digital Video Databases,” International Conference on Pattern Recognition, Brisbane, pp. 916-918 (Aug. 1998); M. K. Mandal et al., “Image Indexing Using Moments and Wavelets,” IEEE Transactions on Consumer Electronics, Vol. 42, No. 3 (Aug. 1996); and S. Pfeiffer et al., “Abstracting Digital Moves Automatically,” Journal on Visual Communications and Image Representation, Vol. 7, No. 4, pp. 345-353 (1996).
The detection of advertising commercials in a video stream is an also active research area. See R. Lienhart et al., “On the Detection and Recognition of Television Commercials,” Proceedings of IEEE International Conference on Multimedia Computing and Systems, pp. 509-516 (1997); and T. McGee et al., “Parsing TV Programs for Identification and Removal of Non-Story Segments,” SPIE Conference on Storage and Retrieval in Image and Video Databases, San Jose (Jan. 1999).
Recognition of text in document images is well known in the art. Document scanners and associated optical character recognition (OCR) software are widely available and well understood. However, detection and recognition of text in video frames presents unique problems and requires a very different approach than does text in printed documents. Text in printed documents is usually restricted to single-color characters on a uniform background (plain paper) and generally requires only a simple thresholding algorithm to separate the text from the background. By contrast, characters in scaled-down video images suffer from a variety of noise components, including uncontrolled illumination conditions. Also, the background frequently moves and text characters may be of different color, sizes and fonts.
The extraction of characters by local thresholding and the detection of image regions containing characters by evaluating gray-level differences between adjacent regions has been described in “Recognizing Characters in Scene Images,” Ohya et al., IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 16, pp. 214-224 (Feb. 1994). Ohya et al. further discloses the merging of detected regions having close proximity and similar gray levels in order to generate character pattern candidates.
Using the spatial context and high contrast characteristics of video text to merge regions with horizontal and vertical edges in close proximity to one another in order to detect text has been described in “Text, Speech, and Vision for Video Segmentation: The Informedia Project,” by A. Hauptmann et al., AAAI Fall 1995 Symposium on Computational Models for Integrating Language and Vision (1995). R. Lienhart and F. Suber discuss a non-linear red, green, and blue (RGB) color system for reducing the number of colors in a video image in “Automatic Text Recognition for Video Indexing,” SPIE Conference on Image and Video Processing (Jan. 1996). A subsequent split-and-merge process produces homogeneous segments having similar color. Lienhart and Suber use various heuristic methods to detect characters in homogenous regions, including foreground. characters, monochrome or rigid characters, size-restricted characters, and characters having high contrast in comparison to surrounding regions.
Using multi-valued image decomposition for locating text and separating images into multiple real foreground and background images is described in “Automatic Text Location in Images and Video Frames,” by A. K. Jain and B. Yu, Proceedings of IEEE Pattern Recognition, pp. 2055-2076, Vol. 31 (Nov. 12, 1998). J-C. Shim et al. describe using a generalized region-labeling algorithm to find homogeneous regions and to segment and extract text in “Automatic Text Extraction from Video for Content-Based Annotation and Retrieval,” Proceedings of the International Conference on Pattern Recognition, pp. 618-620 (1998). Identified foreground images are clustered in order to determine the color and location of text.
Other useful algorithms for character segmentation are described by K. V. Mardia et al. in “A Spatial Thresholding Method for Image Segmentation,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 10, pp. 919-927 (1988), and by A. Perez et al. in “An Iterative Thresholding Method for Image Segmentation,” IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 9, pp. 742-751 (1987).
The prior art text-recognition systems do not take into account, however, the non-semantic attributes of text detected in the content of the video. The prior art systems simply identify the semantic content of the image text and index the video clips based on the semantic content. Other attributes of the image text, such as physical location in the frame, duration, movement, and/or temporal location in a program are ignored. Additionally, no attempt has been made to use video content to identify and edit video clips.
There is therefore a need in the art for improved video processing systems that enable a user to search through an archive of video clips and to selectively save and/or edit all or portions of video clips that contain image text attributes that match image text attributes selected by a user.
SUMMARY OF THE INVENTION
To address the above-discussed deficiencies of the prior art, the present invention discloses a video processing device for searching or filtering video streams for one or more user-selected image text attributes. Generally, “searching” video streams refers to searching in response to user-defined inputs, whereas “filtering” generally refers to an automated process that requires little or no user input. However, in the disclosure, “searching” and “filtering” may be used interchangeably. An image processor detects and extracts image text from frames in video clips, determines the relevant attributes of the extracted image text, and compares the extracted image text attributes and the user-selected image text attributes. If a match occurs, the video processing device may modify, transfer, label or otherwise identify at least a portion of the video stream in accordance with user commands. The video processing device uses the user-selected image text attributes to search through an archive of video clips to 1) locate particular types of events, such as news programs or sports events; 2) locate programs featuring particular persons or groups; 3) locate programs by name; 4) save or remove all or some commercials, and to otherwise sort, edit, and save all of, or portions of, video clips according to image text that appears in the frames of the video clips.
It is a primary object of the present invention to provide, for use in a system capable of analyzing image text in video frame
Agnihotri Lalitha
Dimitrova Nevenka
Elenbaas Jan H.
Do Anh Hong
Goodman Edward W.
Koninklijke Philips Electronics , N.V.
LandOfFree
Method and system for analyzing video content using detected... 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 system for analyzing video content using detected..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for analyzing video content using detected... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3098687