Unfolded convolution for fast feature extraction
First Claim
Patent Images
1. A computer-implemented method for extracting features, comprising:
- one or more processors implementing instructions stored in one or more processor-accessible storage medium for;
receiving a plurality of input features, If, from an input signal, the input features having a height, Iy, and a width, Ix;
unfolding the input features to generate linearized weights, the linearized weights having a height, Ky, and a width, Kx;
determining a horizontal subsampling location, Sx, and a vertical subsampling location, Sy;
concatenating the input features into an input matrix via a processor controllable convolution component, wherein a horizontal component of the input matrix comprises the product of Kx*Ky*If, and a vertical component of the input matrix comprises the product of [((Ix−
Kx+1)+(Sx−
1))/Sx]*[(Iy−
Ky+1)+(Sy−
1))/Sy]; and
multiplying the input matrix by a kernel matrix to generate a matrix product.
2 Assignments
0 Petitions
Accused Products
Abstract
Systems and methods are described that facilitate performing feature extraction across multiple received input features to reduce computational overhead associated with feature processing related to, for instance, optical character recognition. Input feature information can be unfolded and concatenated to generate an aggregated input matrix, which can be convolved with a kernel matrix to produce output feature information for multiple output features concurrently.
-
Citations
13 Claims
-
1. A computer-implemented method for extracting features, comprising:
one or more processors implementing instructions stored in one or more processor-accessible storage medium for; receiving a plurality of input features, If, from an input signal, the input features having a height, Iy, and a width, Ix; unfolding the input features to generate linearized weights, the linearized weights having a height, Ky, and a width, Kx; determining a horizontal subsampling location, Sx, and a vertical subsampling location, Sy; concatenating the input features into an input matrix via a processor controllable convolution component, wherein a horizontal component of the input matrix comprises the product of Kx*Ky*If, and a vertical component of the input matrix comprises the product of [((Ix−
Kx+1)+(Sx−
1))/Sx]*[(Iy−
Ky+1)+(Sy−
1))/Sy]; andmultiplying the input matrix by a kernel matrix to generate a matrix product. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9)
-
10. A computer-implemented system that employs a convolutional neural network to facilitate feature extraction for optical character recognition, comprising:
-
a processor; and a memory into which a plurality of computer-executable instructions are loaded, the plurality of instructions performing a method comprising; receiving a plurality of input features from an input signal, the input features having a height, Iy, and a width, Ix, wherein If is the number of received input features; weighting the input features to generate linearized weights, the linearized weights gleaned from the input features, the linearized weights having a height, Ky, and a width, Kx; unfolding multiple input features; concatenating the unfolded input features to generate an input feature matrix, wherein a horizontal component of the input matrix comprises the product Kx*Ky*If, and wherein a vertical component of the input feature matrix comprises the product of [((Ix−
Kx+1)+(Sx−
1))/Sx]*[(Iy−
Ky+1)+(Sy−
1))/Sy], where Sx is a horizontal subsampling location and Sy is a vertical subsampling location;generating a kernel matrix; and computing a matrix of output features by generating a matrix product of the input feature matrix and the kernel matrix. - View Dependent Claims (11, 12, 13)
-
Specification