Image analysis – Pattern recognition – Feature extraction
Reexamination Certificate
1999-09-03
2003-07-15
Dastouri, Mehrdad (Department: 2623)
Image analysis
Pattern recognition
Feature extraction
C382S240000, C382S260000, C375S240190
Reexamination Certificate
active
06594391
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to techniques for processing images, video and other types of information signals, and more particularly to systems and devices for analyzing and otherwise processing information signals which can be characterized as including textures.
BACKGROUND OF THE INVENTION
Flexible manipulation of image textures has become increasingly important in image processing applications such as computer graphics, video editing, photo-journalism, art, fashion, cataloging, retailing, interactive computer-aided design (CAD), geographic data processing, etc. For example, image texture mapping has become an essential part of computer graphics applications in that it greatly improves the rendering quality of a three-dimensional (3D) environment. The texture mapping approach in such applications is often a simple mapping based on texture replication, which is particularly well-suited for real-time processing.
Conventional image texture analysis and processing techniques for use in such applications are described in, e.g., H. Tamura et al., “Textural features corresponding to visual perception,” IEEE Transactions on Systems, Man and Cybernetics, Vol. 8, pp. 460-473, 1982. The Tamura feature set and other conventional texture processing systems generally concentrate on gray-level natural textures. Another example of such an approach, as described in A. R. Rao and G. L. Lohse, “Towards a texture naming system: Identifying relevant dimensions of texture,” Vision Res., Vol. 36, No. 11, pp. 1649-1669, 1996, focuses on how people classify textures in meaningful, hierarchically-structured categories, identifying relevant features used in the perception of gray-level textures. These and other studies of textual perception have thus focused primarily on developing reliable models for texture discrimination, as well as on the detection of perceptual categories to whose variations humans are the most sensitive.
An important issue which arises in many texture processing applications relates to the replicability of the texture. Replicability is clearly a major concern in the above-noted 3D computer graphics texture mapping. As another example, consider an application in which an image in the form of a “swatch” or other type of sample is stored in an electronic database of wallpaper or upholstery patterns. Although it is typical for the sample to represent a small sub-unit of a larger-scale pattern, the sample itself may be too small to contain enough information to regenerate the larger-scale pattern. This poses a problem if, e.g., there is a desire to generate a display representing the overall appearance of a room or other large surface when wallpapered, upholstered, etc. with a selected pattern. It is generally necessary to replicate the sample provided in order to cover the desired surface. However, misleading and even unpleasant distortions may result if the provided sample is in fact not replicable. Similar concerns are present in other applications involving texture replication. Nonetheless, certain corrections can often be made to a non-replicable texture, in order to improve the appearance of the surface over which it has been replicated. It is therefore important to be able to determine whether a particular image texture sample is replicable.
Wavelet transforms have been widely used for description of image textures. Although numerous decomposition schemes have been proposed for the purpose of texture characterization, numerical description is mainly achieved through second or higher order statistics of the wavelet coefficients. Unfortunately, existing wavelet transforms generally do not measure texture quality along the most important perceptual dimensions. In other words, existing wavelet transform techniques as applied to images fail to quantify textures in the way humans typically see them, i.e., as coarse, regular, directional, symmetric, etc.
As is apparent from the above,, a need exists for improved techniques for analyzing, classifying and otherwise processing image textures, and for accurately and efficiently determining the replicability of such textures.
SUMMARY OF THE INVENTION
The invention provides improved techniques for analysis of textures, which involve characterizing the textures along perceptual dimensions such as directionality, symmetry, regularity and type of regularity. In an illustrative embodiment, a wavelet decomposition or other type of filtering operation is first applied to an image or other information signal including a texture, in order to generate oriented subbands. Texture-related features e.g., coarse features, vertical edges, horizontal edges, corners, etc. are extracted from the subbands. The extracted features are then used to characterize the above-noted perceptual dimensions of the texture. A classification process may subsequently be used to classify the texture as directional, symmetric, regular and/or having a particular regularity type, e.g., a small, medium or large regularity, based on comparison of the determined dimensions for the texture with established thresholds. Advantageously, the classification process can also be used to determine the replicability of the texture.
The techniques of the present invention may be implemented in one or more software programs running on a personal computer, server, workstation, microcomputer, mainframe computer or any other type of programmable digital processor. The invention substantially improves the analysis, replicability determination and other processing of information relating to texture. In addition, the invention can be configured to provide real-time processing of texture information. These and other features and advantages of the present invention will become more apparent from the accompanying drawings and the following detailed description.
REFERENCES:
patent: 5561724 (1996-10-01), Kido et al.
patent: 6222941 (2001-04-01), Zandi et al.
S.G. Mallat, “A Theory for Multiresolution Signal Decomposition: the Wavelet Representation”, Jul. 1989; IEEE Transaction on PAMI, ISSN: 0162-8828, vol. II, No. 7, pp. 674-693.*
Rao et al, “Towards a Texture Naming System: Identifying Relevant Dimensions of Texture”, Oct. 1993; IEEE Conference on Visualization, ISBN: 0-8186-3940-7, pp. 220-227.*
D. Schweitzer, “Artificial Texturing: An Aid to Surface Visualization”, Jul. 1983; Computer Graphics, ISBN: 0-89791-109-1, vol. 17, No. 3, pp. 23-29.*
L. Taycher, “Image Feature Extraction Subsystem of the ImageRover W W W Image Search System”, Aug. 1997; Master Project at Boston University, pp. 1-22.*
W. A. Barrett, “A Survey of Face Recognition Algorithms and Testing Results”, Nov. 1997; IEEE Conference on SSC, ISBN: 0-8186-8316-3, vol. 1, pp. 301-305.*
H. Tamura et al., “Textural Features Correspondng to Visual Perception,” IEEE Transactions on Systems, Man and Cybernetics, vol. SMC-8, No. 6, pp. 460-473, Jun. 1978.
T.V. Papathomas, “A Human Vision Based Computational Model for Chromatic Texture Segregation,” IEEE Transactions on Systems, Man, and Cybernetics-Part B: vol. 27., No. 3, pp. 428-439, Jun. 1997.
W.Y. Ma et al., “Tools for Texture/Color Based Search of Images,” SPIE, vol. 3016, pp. 496-505, 1997.
R. Manmatha et al., “A Syntactic Characterization of Appearance and its Application to Image Retrieval,” SPIE, vol. 3016, pp. 484-495, 1997.
T. Frese et al., “A Methodology for Designing Image Similarity Metrics Based on Human Visual System Models,” In. Proc. of SPIE, vol. 3016, pp. 472-479, 1997.
A.R. Rao et al., “Towards a Texture Naming system: Identifying Relevant Dimensions of Texture,” Vision Res., vol. 36, No. 11, pp. 1650-1663, 1996.
Mojsilovic Aleksandra
Quadranti Laurent Lorenzo Balmelli
Dastouri Mehrdad
Lucent Technologies - Inc.
Ryan & Mason & Lewis, LLP
LandOfFree
Method and apparatus for texture analysis and replicability... 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 texture analysis and replicability..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for texture analysis and replicability... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3092618