Image Matching Using Subspace-Based Discrete Transform Encoded Local Binary Patterns
AssigneeCarnegie Mellon University
InventorsMarios Savvides, Felix Juefei-Xu
Life of this patent
Estimating the term…
Abstract
Determining a match between the subjects of first and second images as a function of decimal-number representations of regions of the first and second images. The decimal-number representations are generated by performing discrete transforms on the regions so as to obtain discrete-transform coefficients, performing local-bit-pattern encoding of the coefficients to create data streams, and converting the data streams to decimal numbers. In one embodiment, the first and second images depict periocular facial regions, and the disclosed techniques can be used for face recognition, even where a small portion of a person's face is captured in an image. Subspace modeling may be used to improve accuracy.
Claims22
A method of matching first features of first image data with second features of second image data, comprising:
- receiving, by a feature-matching system, the first image data;
- applying, by the feature-matching system, a discrete transform to a region of the first image data containing the first features so as to generate a first set of transform coefficients;
- executing, by the feature-matching system, a local binary pattern (LBP) encoding algorithm to encode the first set of transform coefficients into a set of first encoded data streams;
- converting, by the feature-matching system, each of the first encoded data streams into a decimal number so as to create a first decimal-number representation of the region of the first image data;
- and
- executing, by the feature-matching system, a matching algorithm for determining a match as a function of the first decimal number representation and a second decimal representation of a region of the second image data containing the second features.
A machine-readable storage medium containing machine-executable instructions for performing a method of matching first features of first image data with second features of second image data, said machine-executable instructions comprising:
- a first set of machine-executable instructions for receiving the first image data;
- a second set of machine-executable instructions for applying a discrete transform to a region of the first image data containing the first features so as to generate a first set of transform coefficients;
- a third set of machine-executable instructions for executing a local binary pattern (LBP) encoding algorithm to encode the first set of transform coefficients into a set of first encoded data streams;
- a fourth set of machine-executable instructions for converting each of the first encoded data streams into a decimal number so as to create a first decimal-number representation of the region of the first image data;
- and
- a fifth set of machine-executable instructions for executing a matching algorithm for determining a match as a function of the first decimal number representation and a second decimal representation of a region of the second image data containing the second features.
Dependencies are inferred from the claim text.