System and method for automatic linguistic indexing of images by a statistical modeling approach
First Claim
1. A method for automatic linguistic indexing of images, said method comprising the steps of:
- establishing a database of statistical models using a statistical modeling process, wherein the database is accessible by the computer, and each of the statistical models represents a predetermined semantic category;
associating a set of index terms with each statistical model in the database, wherein the set of index terms provides a description for the predetermined semantic category;
extracting a plurality of feature vectors from an image to be indexed;
statistically comparing the plurality of feature vectors extracted from the image to be indexed to the statistical models in the database;
determining a set of statistical models from the database that are statistically similar to the plurality of feature vectors extracted from the image to be indexed; and
extracting a set of statistically significant index terms from the descriptions of the set of statistical models and using the set of statistically significant indexed terms to index the image.
1 Assignment
0 Petitions
Accused Products
Abstract
The present invention provides a statistical modeling approach to automatic linguistic indexing of photographic images. The invention uses categorized images to train a dictionary of hundreds of statistical models each representing a concept. Images of any given concept are regarded as instances of a stochastic process that characterizes the concept. To measure the extent of association between an image and a textual description associated with a predefined concept, the likelihood of the occurrence of the image based on the characterizing stochastic process is computed. A high likelihood indicates a strong association between the textual description and the image. The invention utilizes two-dimensional multi-resolution hidden Markov models that demonstrate accuracy and high potential in linguistic indexing of photographic images.
246 Citations
26 Claims
-
1. A method for automatic linguistic indexing of images, said method comprising the steps of:
-
establishing a database of statistical models using a statistical modeling process, wherein the database is accessible by the computer, and each of the statistical models represents a predetermined semantic category; associating a set of index terms with each statistical model in the database, wherein the set of index terms provides a description for the predetermined semantic category; extracting a plurality of feature vectors from an image to be indexed; statistically comparing the plurality of feature vectors extracted from the image to be indexed to the statistical models in the database; determining a set of statistical models from the database that are statistically similar to the plurality of feature vectors extracted from the image to be indexed; and extracting a set of statistically significant index terms from the descriptions of the set of statistical models and using the set of statistically significant indexed terms to index the image. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12)
-
-
13. A method of automatic linguistic indexing of images using a computer system, said method comprising the steps of:
-
generating a training image database having a plurality of training images wherein the plurality of training images represent at least one semantic category; assigning a textual description to the at least one semantic category in the training database; extracting a plurality of feature vectors from the training images using a statistical modeling process; generating statistical models using the extracted feature vectors, wherein the statistical models are associated with portions of the textual description assigned to the at least one category and including a plurality of paths, each of the plurality of paths representing a variation of the training images; storing the statistical models; and using the stored models to index and assign a textual description to an image by statistically comparing the models with a plurality of features extracted from the image to be indexed to determine the statistical similarity between the image and each model. - View Dependent Claims (14, 15, 16, 17, 18)
-
-
19. A computer system for use in automatic linguistic indexing of images, said system comprising:
-
a computer operative to receive an image to be indexed and assigned a textual description; a plurality of different semantic categories disposed on said computer; a statistical modeling algorithm operative to construct a statistical model representative of each one of said plurality of different semantic categories; a database in communication with said computer for storing said plurality of different semantic categories and said statistical models, each of said different semantic categories and said statistical models having a predetermined textual description associated therewith; a feature extraction algorithm operative to extract a plurality of feature vectors from the image to be indexed; a feature comparison algorithm operative to statistically compare said plurality of feature vectors extracted from the image with each of said statistical models to determine statistical similarity between the image and each of said statistical models; and a text assigning algorithm operative to extract a set of statistically significant index terms from said predetermined textual descriptions associated with said statistical models wherein said set of index terms provide the textual description for the image to be indexed. - View Dependent Claims (20, 21, 22, 23, 24, 25, 26)
-
Specification