Dimensionality Reduction of Neural Networks Intermediate Feature Maps Using Two-Dimensional Principal Component Analysis

← USPTO Patent Applications

DIMENSIONALITY REDUCTION OF NEURAL NETWORKS INTERMEDIA FEATURE MAPS USING TWO-DIMENSIONAL PRINCIPAL COMPONENT ANALYSIS

Application US20260099700A1 Kind: A1 Apr 09, 2026

Inventors

Homayun AFRABANDPEY, Alireza AMINLOU, Hamed REZAZADEGAN TAVAKOLI, Honglei ZHANG, Miska Matias HANNUKSELA

Abstract

The embodiments concern a method comprising: decoding, from or along a bitstream, a mean matrix, a principal components matrix, and a row projection matrix; wherein the mean matrix corresponds to a mean of training data matrices, wherein the training data matrices are respective slices of at least one input tensor along a channel dimension of the at least one input tensor; wherein an original input matrix is a slice of the at least one input tensor along the channel dimension of the at least one input tensor, wherein the original input matrix has dimensions comprising at least a height and a width; wherein the at least one input tensor corresponds to at least one input image; wherein the row projection matrix comprises a concatenation of row projection vectors; and reconstructing the original input matrix by adding the mean matrix to a product comprising a multiplication of the principal components matrix with a transpose of the row projection matrix. The embodiments also concern technical equipment for implementing the method.

CPC Classifications

G06N 3/0455 G06N 3/044

Filing Date

2025-10-06

Application No.

19350801

View original document →