Object Recognition Method and Apparatus Using Convolutional Neural Networks

ChangeBridge: Patent Grants - AI & Computing (G06N)

Published March 31st, 2026

Detected April 1st, 2026

Summary

USPTO granted Patent US12592069B2 to Shandong Yingxin Computer Technologies Co., Ltd. on March 31, 2026, covering an object recognition method using convolutional neural networks. The patent describes a technical process for extracting and analyzing feature information through convolutional layers, pooling layers, and fully connected layers to identify objects and compile recognition statistics.

View original document View source feed page

What changed

USPTO issued Patent US12592069B2 for an AI-based object recognition method and apparatus utilizing convolutional neural networks. The patent covers processing steps including feature extraction via convolutional layers, dimension reduction through pooling, and comparative analysis using fully connected layers against pre-stored queue information to output recognition results. The patent has 18 claims and was filed on January 28, 2022.

This is a routine patent grant notice with no compliance obligations or deadlines for other entities. Technology companies developing similar AI/ML systems for object recognition should review this patent to assess potential intellectual property considerations and freedom-to-operate implications for their own products.

Source document (simplified)

← USPTO Patent Grants

Object recognition method and apparatus, and device and medium

Grant US12592069B2 Kind: B2 Mar 31, 2026

Assignee

SHANDONG YINGXIN COMPUTER TECHNOLOGIES CO., LTD.

Inventors

Fancheng Meng, Hongliang Wang, Qi Mou

Abstract

Disclosed are an object recognition method and apparatus, and a device and a medium. The method comprises: by a convolutional layer in a convolutional neural network, extracting pre-selected feature information of objects to be recognized, and performing dimension reduction processing on the pre-selected feature information by using a pooling layer, to obtain feature information; processing the feature information by a fully connected layer, to obtain processed feature information, extracting a target feature value from the processed feature information, and compiling statistics on the number of objects to be recognized; by the fully connected layer and according to the target feature value, determining corresponding target queue information from pre-stored queue information; and by the fully connected layer, performing a comparative analysis on the processed feature information and reference feature information of each sample object in a queue corresponding to the target queue information, and outputting a corresponding recognition result.