Fast pattern classification based on a sparse transform
First Claim
1. A method for determining a feature in a subject image, comprising:
- performing a sparse transform on a subject image, wherein the sparse transform is selected from a group comprising a wavelet transform and a Fourier transform, to determine a sparse transformed subject image that represents the subject image with a few significant coefficients compared to a number of values in the subject image at each of a plurality of orientations;
receiving a plurality of number I of patch functions that each is based on a portion of a sparse transformed image at corresponding orientations for a selected image of a training set of images, wherein the plurality of patch functions represent learned features in the training set; and
sending data that indicates a feature in the subject image based on the transformed subject image and the plurality of patch functions.
1 Assignment
0 Petitions
Accused Products
Abstract
Techniques for determining a feature in an image or soundtrack of one or more dimensions include receiving a subject image. A sparse transformed subject image is determined, which represents the subject image with a few significant coefficients compared to a number of values in the subject image. Multiple patch functions are received, which are based on a portion of a sparse transformed image for each of a training set of images and which represent learned features in the training set. A feature is determined to be in the subject image based on the transformed subject image and the plurality of patch functions. In various embodiments, a wavelet transformation or audio spectrogram is performed to produce the sparse transformed images. In some embodiments, the feature in the subject is determined regardless of feature location or size or orientation in the subject image.
41 Citations
23 Claims
-
1. A method for determining a feature in a subject image, comprising:
-
performing a sparse transform on a subject image, wherein the sparse transform is selected from a group comprising a wavelet transform and a Fourier transform, to determine a sparse transformed subject image that represents the subject image with a few significant coefficients compared to a number of values in the subject image at each of a plurality of orientations; receiving a plurality of number I of patch functions that each is based on a portion of a sparse transformed image at corresponding orientations for a selected image of a training set of images, wherein the plurality of patch functions represent learned features in the training set; and sending data that indicates a feature in the subject image based on the transformed subject image and the plurality of patch functions.
-
-
2. A non-transitory computer-readable medium carrying one or more sequences of instructions for determining a feature in a subject image, wherein execution of the one or more sequences of instructions by one or more processors causes the one or more processors to perform:
-
performing a sparse transform on a subject image, wherein the sparse transform is selected from a group comprising a wavelet transform and a Fourier transform, to determine a sparse transformed subject image that represents the subject image with a few significant coefficients compared to a number of values in the subject image at each of a plurality of orientations; receiving a plurality of number I of patch functions that each is based on a portion of a sparse transformed image at corresponding orientations for a selected image of a training set of images, wherein the plurality of patch functions represent learned features in the training set; and determining a feature in the subject image based on the transformed subject image and the plurality of patch functions. - View Dependent Claims (3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14)
-
-
15. An apparatus for determining a feature in an image, comprising:
-
means for performing a sparse transform on a subject image, wherein the sparse transform is selected from a group comprising a wavelet transform and a Fourier transform, to determine a sparse transformed subject image that represents the subject image with a few significant coefficients compared to a number of values in the subject image at each of a plurality of orientations; means for-receiving a plurality of number I of patch functions that each is based on a portion of a sparse transformed image at corresponding orientations for a selected image of a training set of images, wherein the plurality of patch functions represent learned features in the training set; and means for determining a feature in the subject image based on the transformed subject image and the plurality of patch functions.
-
-
16. An apparatus for determining a feature in an image, comprising logic encoded in a tangible medium and configured to:
-
perform a sparse transform on a subject image, wherein the sparse transform is selected from a group comprising a wavelet transform and a Fourier transform, to determine a sparse transformed subject image that represents the subject image with a few significant coefficients compared to a number of values in the subject image at each of a plurality of orientations; receive a plurality of number I of patch functions that each is based on a portion of a sparse transformed image at corresponding orientations for a selected image of a training set of images, wherein the plurality of patch functions represent learned features in the training set; and determine a feature in the subject image based on the transformed subject image and the plurality of patch functions. - View Dependent Claims (17, 18, 19, 20)
-
-
21. A method for determining a feature in a subject image, comprising:
-
performing a sparse transform on a subject image, wherein the sparse transform is selected from a group comprising a wavelet transform and a Fourier transform, to determine a sparse transformed subject image that represents the subject image with a few significant coefficients compared to a number of values in the subject image at each of one or more orientations; receiving a plurality of number I of patch functions that each is based on a random portion of a sparse transformed image at corresponding orientations for a randomly selected image of a training set of images, wherein the plurality of patch functions represent learned features in the training set; and sending data that indicates a feature in the subject image based on the transformed subject image and the plurality of patch functions.
-
-
22. A non-transitory computer-readable medium carrying one or more sequences of instructions for determining a feature in a subject image, wherein execution of the one or more sequences of instructions by one or more processors causes the one or more processors to perform:
-
performing a sparse transform on a subject image, wherein the sparse transform is selected from a group comprising a wavelet transform and a Fourier transform, to determine a sparse transformed subject image that represents the subject image with a few significant coefficients compared to a number of values in the subject image at each of one or more orientations; receiving a plurality of number I of patch functions that each is based on a random portion of a sparse transformed image at corresponding orientations for a randomly selected image of a training set of images, wherein the plurality of patch functions represent learned features in the training set; and determining a feature in the subject image based on the transformed subject image and the plurality of patch functions.
-
-
23. An apparatus for determining a feature in an image, comprising logic encoded in a tangible medium and configured to:
-
perform a sparse transform on a subject image, wherein the sparse transform is selected from a group comprising a wavelet transform and a Fourier transform, to determine a sparse transformed subject image that represents the subject image with a few significant coefficients compared to a number of values in the subject image at each of one or more orientations; receive a plurality of number I of patch functions that each is based on a random portion of a sparse transformed image at corresponding orientations for a randomly selected image of a training set of images, wherein the plurality of patch functions represent learned features in the training set; and determine a feature in the subject image based on the transformed subject image and the plurality of patch functions.
-
Specification