Yahoo! Inc.

Shesha Shah

Sriram Jayaprakash Sathish

A content-based approach is provided for detecting near duplicate images. The approach generally involves analyzing the content of images to be compared and generating color and texture histogram descriptor data for each image. The images may then be compared based upon the color and texture histogram descriptor data to determine whether the images are near duplicate images. Content-based image signature data may also be generated for each of the images based upon the color and texture histogram descriptor data. The image signature data may then be compared to determine whether the corresponding images are near duplicate images.

Approach for near duplicate image detection

A method and apparatus for, in response to receiving a request, retrieving digital images that are associated with one or more keywords and which are similar to a digital image is provided. The request may be processed using a keywords index and a content index, and optionally, a session index that references data about what digital images users have viewed together in a session. A category associated with the identified digital image may also be used in processing the request, e.g., the category associated with the identified digital image may change the weight placed on various content attributes used in comparing the content of the identified digital image to other digital images. After the set of requested digital images are determined, request data, that identifies the set of requested digital images, may be sent to the requestor.

Method and mechanism for retrieving images

A method and apparatus for processing image data is provided. Color image data is processed to generate gray scale image data, which may comprise a set of data values. Each data value, of the set of data values, may identify a gray scale value. A frequency distribution that indicates a frequency of occurrence of each gray scale value identified by the set of data values is determined. The uniformity of the frequency distribution is determined. A determination of how close the frequency determination corresponds to a predetermined distribution is made. The color image data is classified as a particular type of image data based on (a) whether the frequency determination corresponds to the predetermined distribution by a first threshold, and (b) whether the uniformity of the frequency distribution exceeds a second threshold.

Method and mechanism for processing image data

A method and apparatus for comparing content attributes, e.g., texture, of one digital image a second digital image. An image operation is performed on gray scale image data to identify intensity transitions of a first digital image. The gray scale image data comprises a set of data values that each identifies a gray scale value. For at least a subset of the set of data values, a slope value for an intensity transition associated with the data value is determined. The slope values are converted to possible slope values of a bounded set of possible slope values. First slope data that describes a frequency of occurrence, across the subset of data values, of each possible slope value is generated. The first slope data is compared to second slope data associated with the second digital image to determine how similar the first digital image is to the second digital image.

Growing community of inventors

Bengaluru, India

Shesha Shah