A fast face retrieval method based on depth learning

A technology of deep learning and retrieval algorithms, applied in still image data retrieval, instruments, biological neural network models, etc., can solve problems such as low efficiency and limited application scenarios of verification algorithms, achieve application stability, improve retrieval efficiency, and robustness Good results

Active Publication Date: 2019-01-18
UNIV OF ELECTRONICS SCI & TECH OF CHINA
View PDF8 Cites 8 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In practical applications, the number of samples is generally large and will change frequently (such as the ID card database of citizens), so for the face recognition problem, most researchers focus on verification, that is, comparing whether two pictures belong to the same individual However, the application scenarios of the verification algorithm are limited, and the retrieval of massive databases can only be done through one-to-one comparison, which is relatively inefficient. Therefore, how to quickly retrieve faces in massive databases is a key research direction of face recognition

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 fast face retrieval method based on depth learning
  • A fast face retrieval method based on depth learning
  • A fast face retrieval method based on depth learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0050] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0051] like figure 1 As shown, a fast face retrieval method based on deep learning includes the following steps:

[0052] Step 1: If figure 2 As shown, design the convolutional neural network structure and loss function;

[0053] The convolutional neural network uses Inception-ResNet-v2 as the main structure and the Sigmoid function as the activation function. The loss functions include cross-entropy loss function, center distance loss layer loss function and inter-class loss layer loss function;

[0054] (1) Cross entropy loss function:

[0055] The expression of the cross entropy loss function is as follows:

[0056]

[0057] Where a is the actual output of the network, and y is the expected value.

[0058] (2) Center distance loss layer loss function:

[0059] The expression of the center distance loss function is as follows:

[0...

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 discloses a fast face retrieval method based on depth learning, belonging to the application field of computer vision object recognition. The invention mainly utilizes the depth neural network and extracts the face binary features, and combines the local sensitive hashing (LSH) nearest neighbor retrieval algorithm to propose a fast and efficient face retrieval algorithm. Firstly, thedepth neural network is trained by a large number of calibrated face data, and then all the images in the database are substituted into the neural network for calculation to obtain the correspondingbinary feature library, and then the hash function is designed to construct the hash structure. For the images to be retrieved, the images to be retrieved is also substituted into the neural network to calculate the binary features, and then the LSH algorithm is used to complete the retrieval. Through a large number of database tests, it is found that this method has high retrieval accuracy, highspeed, and good robustness in the engineering environment.

Description

technical field [0001] The invention relates to the application field of computer vision target retrieval, in particular to a fast face retrieval method based on deep learning in a large data set. Background technique [0002] The face recognition system takes face recognition technology as the core. It is an emerging biometric technology and a high-tech technology in the international scientific and technological field. It widely uses regional feature analysis algorithms, integrates computer image processing technology and biostatistics principles, uses computer image processing technology to extract portrait feature points from videos, and uses biostatistics principles to analyze and establish mathematical models. Prospects. Face retrieval technology is widely used in many fields and scenarios such as criminal investigation, monitoring, and public security, and has very high research value. As the complexity of application scenarios continues to increase, the scale of th...

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 Applications(China)
IPC IPC(8): G06F16/53G06K9/00G06K9/62G06N3/04
CPCG06V40/172G06V40/168G06N3/045G06F18/214
Inventor 陶冰洁王酉祥马泸敏廖龙
Owner UNIV OF ELECTRONICS SCI & TECH OF CHINA
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap