A face detection method, device, and medium based on dual-stream feature extraction

By combining YOLOv11 and the improved MViTv2 branch, the complementarity of global and local features is achieved, which solves the detection accuracy problem in dense and occluded scenes and improves the stability and adaptability of the detection system.

CN121438374BActive Publication Date: 2026-06-30SHANDONG RUICHUANG EDUCATION TECH CO LTD
1 Cites 0 Cited by

Patent Information

Application Number
CN202511662718.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-11-13
Publication Date
2026-06-30
Estimated Expiration
2045-11-13

AI Technical Summary

Technical Problem

In dense and heavily occluded scenarios, the existing YOLOv1 network struggles to effectively identify heavily occluded targets and large targets, resulting in a decrease in detector recall and failing to meet actual deployment requirements.

Method used

A face detection method based on dual-stream feature extraction is adopted, which combines the local feature extraction capability of YOLOv11 with the global modeling capability of the optimized MViTv2. The local and global features are fused through a feature fusion module, and an improved Dynamic ATSS label allocation strategy and soft label supervision mechanism are introduced to optimize the MViTv2 branch and enhance the detection capability of targets of different sizes.

Benefits of technology

It significantly improves detection accuracy in scenarios with severe occlusion or dense multi-target environments, solves the problem of insufficient local receptive field in traditional convolutional networks, enhances the stability and reliability of the detection system, and adapts to the detection needs of targets of different sizes.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

This invention provides a face detection method, apparatus, and medium based on dual-stream feature extraction, belonging to the field of computer vision technology. The method includes: acquiring face images to establish a dataset; preprocessing the dataset and dividing it into training and testing sets; constructing a face detection model, which includes a YOLOv11 branch, an MViTv2 branch, a feature fusion module, and a detection head. The YOLOv11 branch uses the backbone network of the YOLOv11 network to extract local features, the MViTv2 branch uses an improved MViTv2 model to extract global features, the feature fusion module fuses local and global features at corresponding scales, and the detection head outputs detection results based on the fused features; training the face detection model using the training set data; acquiring face images to be detected and inputting them into the trained face detection model to obtain detection results. This invention effectively alleviates the difficulty of matching overlapping and occluded targets in dense scenes, further improving detection performance.
Need to check novelty before this filing date? Find Prior Art