Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A face feature comparison method based on cuda technology

A feature comparison and face feature technology, applied in the field of biometrics, can solve problems such as increased hardware costs and low concurrent efficiency of face features, and achieve the effects of improving efficiency, improving concurrent efficiency, and increasing speed

Active Publication Date: 2021-09-24
GOSUNCN TECH GRP
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to provide a face feature comparison method based on cuda technology, to solve the problem that the concurrency efficiency of face feature comparison in the prior art is not high, and it is necessary to increase the comparison server for horizontal expansion, resulting in increased hardware costs question

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A face feature comparison method based on cuda technology
  • A face feature comparison method based on cuda technology
  • A face feature comparison method based on cuda technology

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0031] The present invention will be further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings.

[0032] This solution needs to pre-load all the face features of the face feature library into the global memory of the GPU, and ensure that the memory blocks are continuous and aligned.

[0033] Based on the hardware architecture of cuda (Compute United Device Architecture) and the memory access mechanism, by calculating one feature per thread block, it solves the combined access of GPU global memory, reduces the number of thread accesses to global memory, and maximizes the bandwidth of global memory. optimization. In order to improve the comparison speed of a single feature comparison request, the specific steps are as follows figure 1 shown, including:

[0034] Step a, obtaining effective features; effective features refer to facial features that can be effectively detected and recognized;

[0035] Step b. Each ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention belongs to the field of biometric identification, and specifically relates to a face feature comparison method based on cuda technology. The method includes: firstly, based on the hardware architecture and memory access mechanism of cuda, by modifying the way the thread accesses the global memory, the previous one A thread’s access to a feature block becomes a thread’s access to a byte in a feature block, which realizes combined access to global memory, reduces the number of threads’ access to global memory, and improves the speed of a single facial feature comparison; Each facial feature comparison request is cached, and all comparison requests are combined for feature comparison according to predetermined rules; finally, the comparison results are split into independent results according to predetermined rules and reported to users, thereby improving human performance. Concurrent efficiency of face feature comparison. Under the condition of the same face feature library capacity and concurrent requests, this solution can effectively reduce server resources and save hardware costs.

Description

technical field [0001] The invention relates to the technical field of biometric identification, in particular to a method for comparing human face features based on cuda technology. Background technique [0002] At present, the biometric technology based on facial features is mainly based on the effective comparison of facial features to complete the face recognition process. By extracting the features of the face in the current image and comparing them one by one with the massive face features in the preset face feature library, the comparison similarity scores are obtained respectively, and then the similarity scores are sorted in descending order. Finally, a series of faces with the highest similarity meeting the threshold is retained as the result for output. [0003] At present, the construction of safe cities across the country is in full swing, and a large amount of preprocessed face feature data has also been produced. Therefore, the effective face feature comparis...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/00
CPCG06V40/172G06V40/168
Inventor 关喜记江盛欣劳定雄洪曙光黄仝宇汪刚宋一兵侯玉清刘双广
Owner GOSUNCN TECH GRP
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com