US 2025/0061577 A1 · Computer Vision Systems and Methods for End-to-End Training of Convolutional Neural Networks Using Differentiable Dual-Decomposition Techniques

Abstract

Computer vision systems and methods for end-to end training of neural networks are provided. The system generates a fixed point algorithm for dual-decomposition of a maximum-a-posteriori inference problem and trains the convolutional neural network and a conditional random field with the fixed point algorithm and a plurality of images of a dataset to learn to perform semantic image segmentation. The system can segment an attribute of an image of the dataset by the trained neural network and the conditional random field.

Claims18

Dependent claims

A computer vision system for end-to-end training of a neural network, comprising:

a memory;
and
a processor in communication with the memory, the processor:
implementing an algorithm for decomposing an inference problem, training a neural network to perform semantic image segmentation by applying the algorithm to training input data;
and
processing one or more images to segment an attribute of the one or more images using the trained neural network.

3 more elements

Dependencies are inferred from the claim text.

Facts

Application

18/903,493Filed 1 Oct 2024

Priority

13 Dec 2019

Published

20 Feb 2025

Term ends

Not estimatedThe source records don’t support an estimate.

Classification27 codes

G06T 7/10
Image analysis: segmentation; Edge detection
G06F 18/2163
Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation: partitioning the feature space
G06F 18/29
Analysing: graphical models, e.g. Bayesian networks
G06F 18/295
Graphical models, e.g. Bayesian networks: markov models or related models, e.g. semi-Markov models; Markov random fields; Networks embedding…

Family

Application62/947,874Us provisional application · filed 13 Dec 2019

Application17/121,257Continuation · filed 14 Dec 2020

Application18/903,493Continuation

US 12,106,481 B2Related patent

US 2021/0182675 A1Published application

US 2025/0061577 A1This patent

From the US filing history. Foreign family members aren’t listed.

Computer Vision Systems and Methods for End-to-End Training of Convolutional Neural Networks Using Differentiable Dual-Decomposition Techniques

Life of this patent

Computer Vision Systems and Methods for End-to-End Training of Convolutional Neural Networks Using Differentiable Dual-Decomposition Techniques

Life of this patent

Abstract

Claims18