Loading patent…
AssigneeInternational Business Machines Corporation
InventorsAleksandra Mojsilovic, Bernice Rogowitz, Jose Gomes
Estimating the term…
A method and system for determining image similarity is based on the semantic meaning of images. The method includes deriving a plurality of semantic categories for representing important semantic cues in images, where each semantic category is modeled through a combination of perceptual features that define the semantics of that category and that discriminate that category from other categories; for each semantic category, forming a set of the perceptual features comprising required features and frequently occurring features; comparing an image to said semantic categories; and classifying said image as belonging to one of said semantic categories if all of the required features and at least one of the frequently occurring features for that semantic category are present in said image. A database contains image information, where the image information includes at least one of already classified images, network locations of already classified images and documents containing already classified images. The database is searched for images matching an input query, that can comprise an image, text, or both. The images stored in or referenced by the database may be obtained at least in part through the Internet, such as by the activity of an automated web crawler. In the presently preferred embodiment the images are medical images, and the medical database may be searched for at least one image that satisfies thresholds established by a search query. Also disclosed is a method and system for searching the Internet. The method includes (a) providing at least one server accessible by a plurality of clients; (b) downloading a program to a client, where the program includes an Internet searching program operable for locating at least one type of data; and (c) executing the downloaded program on the client, and returning to the server URLs corresponding to located data.
A computer implemented method for characterizing images and determining image similarity based on semantic meaning of images, comprising:
A data processing system comprising a data processor, a graphical user interface and a memory that stores a database containing image information, where the image information comprises at least one of already classified images, network locations of already classified images and documents containing already classified images, said data processor operating in accordance with a stored program for classifying images by deriving a plurality of semantic categories for representing important semantic cues in images, where each semantic category is modeled through a combination of perceptual features that define the semantics of that category and that discriminate that category from other categories;
A computer program embodied on a computer readable media for directing a computer to execute a method for processing digitally represented images, comprising program instructions for classifying images by deriving a plurality of semantic categories for representing important semantic cues in images, where each semantic category is modeled through a combination of perceptual features that define the semantics of that category and that discriminate that category from other categories, where for each semantic category said program instructions form a set of the perceptual features comprising required features and frequently occurring features, compares an image to said semantic categories and classifies said image as belonging to one of said semantic categories if all of the required features and at least one of the frequently occurring features for that semantic category are present in said image.
An Internet searching tool, comprising:
A method for searching the Internet, comprising:
A system for obtaining information from the Internet, comprising at least one server accessible by a plurality of clients, said server operable for downloading a program to a client, said program comprising an Internet searching program operable for locating at least one type of data, said downloaded program being executable by said client and returning to said server URLs corresponding to located data, said server further comprising a database, searchable by the clients, for storing information corresponding to information returned by said downloaded program.
Dependencies are inferred from the claim text.
Drawings are temporarily unavailable. Open the PDF at USPTO
From the US filing history. Foreign family members aren’t listed.
Backward citations are indexed for grants only.
ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or…: for…
Of still image data: retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
Using metadata automatically derived from the content: using colour
Using metadata automatically derived from the content: using shape and object relationship
Term end not estimated
These teachings relate generally to database management methodologies and, more specifically, the teachings in accordance with this invention relate to methods and apparatus for managing and operating with a database that contains a set of digitally represented images.
The flexible retrieval from, manipulation of, and navigation through image databases has become an important problem in the database management arts, as it has applications in video editing, photojournalism, art, fashion, cataloguing, retailing, interactive computer aided design (CAD), geographic data processing and so forth.
An early content-based retrieval (CBR) system is one known as ART MUSEUM. Reference in this regard can be made to K. Hirata and T. Katzo, “Query by visual example, content based image retrieval”, in Advances in Database Technology—EDBT'92, A. Pirotte, C. Delobel, and G. Gottlob, Eds., Lecture Notes in Computer Science, vol. 580, 1992. In this particular CBR the retrieval of image data is based entirely on edge features. An early commercial content-based image search engine that had profound effects on later systems was one known as QBIC. Reference in this regard can be had to W. Niblack, R. Berber, W. Equitz, M. Flickner, E. Glasman, D. Petkovic, and P. Yanker, “The QBIC project: Querying images by content using color, texture and shape”, in Proc. SPIE Storage and Retrieval for Image and Video Data Bases, pp.172-187, 1994. For color representation this system uses a k-element histogram and average of (R,G,B), (Y,i,q), and (L,a,b) coordinates, whereas for the description of texture it implements the feature set of Tamura (see H. Tamura, S. Mori, and T. Yamawaki, “Textural features corresponding to visual perception”, IEEE Transactions Systems, Man and Cybernetics, vol. 8, pp. 460-473, 1982.) In a similar fashion, color, texture and shape are supported as a set of interactive tools for browsing and searching images in the Photobook system developed at the MIT Media Lab, as described by A. Pentland, R. W. Picard, and S. Sclaroff, “Photobook: Content-based manipulation of image databases”, International Journal of Computer Vision, vol. 18, no. 3, pp. 233-254, 1996. In addition to these elementary features, systems such as VisualSeek (see J. R. Smith, and S. Chang, “VisualSeek: A fully automated content-based query system”, in Proc. ACM Multimedia 96, pp. 87-98, 1996), Netra (see W. Y. Ma, and B. S. Manjunath, “Netra: A toolbox for navigating large image databases” in Proc. IEEE Int. Conf. on Image Processing, vol. 1, pp. 568-571, 1997) and Virage (see A. Gupta, and R. Jain, “Visual information retrieval”, Communications of the ACM, vol. 40, no. 5, pp. 70-79, 1997) support queries based on spatial relationships and color layout. Moreover, in the Virage system, users can select a combination of implemented features by adjusting weights according to their own “perception”. This paradigm is also supported in the RetrievalWare search engine (see J. Dowe “Content based retrieval in multimedia imaging”, in Proc. SPIE Storage and Retrieval for Image and Video Databases, 1993.) A different approach to similarity modeling is proposed in the MARS system, as described by Y. Rui, T. S. Huang, and S. Mehrotra, “Content-based image retrieval with relevance feed-back in Mars”, in Proc. IEEE Conf on Image Processing, vol. II, pp. 815-818, 1997. In the MARS system the main focus is not on finding a best representation, but rather on the use of relevance feedback to dynamically adapt multiple visual features to different applications and different users.
The foregoing and other aspects of these teachings are made more evident in the following Detailed Description of the Preferred Embodiments, when read in conjunction with the attached Drawing Figures, wherein:
FIG. 1 is simplified block diagram of a data processing system that is suitable for practicing this invention;
FIG. 2 is a logic flow diagram that illustrates a method for computing a similarity metric between an image x and a semantic category ci;
The teachings of this invention are related to some extent to the subject matter described in copending U.S. patent application Ser. No. 10/033,597, filed Dec. 27, 2001, entitled “Perceptual Method for Browsing, Searching, Querying and Visualizing Collections of Digital Images”, by Aleksandra Mojsilovic and Bernice E. Rogowitz, incorporated by reference herein in its entirety. Prior to describing the teachings of this invention, it will be instructive to review the teachings of the commonly assigned U.S. patent application Ser. No. 10/033,597.
In brief, that invention provides an image processing method and system that is based on human perception, and that extracts semantic information about images. The method allows images to be organized and categorized by semantic content, without requiring key words. The method can enable the development of perceptual front-ends to many image applications. The method is implemented using a set of image processing algorithms that extract visual attributes from images and analyzes them to assign semantic meaning.
A first method assigns semantic meaning to an image, without requiring the use of a costly and labor-intensive step where each image is labeled manually with a key word. A second method enables a user to search, navigate, and browse through a library of images based on semantic categories. These are important advantages when developing user-interfaces, and when developing useful multimedia databases.