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

Dynamic evolution expression recognition method and system

A recognition method and expression recognition technology, applied in neural learning methods, character and pattern recognition, instruments, etc., can solve the problems of reduced recognition accuracy, inability of expression recognition technology to generalize application scenarios, and difficulty in retaining model knowledge and memory.

Pending Publication Date: 2020-12-15
TSINGHUA UNIV
View PDF0 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In this case, the recognition method of dynamically evolving expressions becomes an urgent and fundamental challenge in expression recognition technology.
[0003] However, in the existing technology, when using a fixed-type expression recognition method, when introducing data of a new expression category, it is necessary to discard the original expression recognition model, and then use all the data including the original expression category and the new expression category to retrain. The new expression recognition model leads to high memory and time consumption in the actual application process, which limits the application of expression recognition technology in complex expression categories
The root cause of this problem is that it is difficult to retain the model's knowledge memory of the original expression categories when training the facial expression recognition model with dynamically increased categories, and there is a possibility that the recognition accuracy of the original expression categories will be greatly reduced. Recognition technology cannot be generalized to application scenarios where expression categories increase dynamically

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
  • Dynamic evolution expression recognition method and system
  • Dynamic evolution expression recognition method and system
  • Dynamic evolution expression recognition method and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0075] The following combination Figure 1 to Figure 2 Embodiment 1 of the present application will be described.

[0076] Such as figure 1 As shown, the present embodiment provides a method for recognizing dynamically evolving expressions, including:

[0077] Step 1. According to the dynamic increase of expression categories, build and optimize the past expression memory module;

[0078] In this step 1, specifically include:

[0079] Step 11. According to the memory limit of the computing device and the integrity of the data acquisition authority, set the total size K of the representative expression pictures stored in the past expression memory module. Generally speaking, the memory occupied by K pictures should not exceed 10% of the current device's memory, and it will increase moderately with the increase of the completeness of the data acquisition permission;

[0080] Step 12. After the system initialization phase, according to the s in the initialization 0 Class exp...

Embodiment 2

[0145] Such as image 3 As shown, this embodiment provides. A recognition system 10 for dynamically evolving facial expressions, including: a past facial expression memory module 100, a network model construction module 200 and a recognition result output module 300;

[0146] The past expression memory module 100 is configured to, according to the dynamic increase of the expression category and the neural network model of the expression recognition, extract the features corresponding to all objects of the new category of expressions, construct a feature database, and finally optimize and update the past expression memory module;

[0147] Further, the past expression memory module 100 specifically includes: an image acquisition unit 101, a feature extraction unit 102, an arithmetic mean unit 103, a sequence generation unit 104 and a sequence adjustment unit 105; the image acquisition unit 101 is used to obtain the first occurrence The picture objects belonging to the same categ...

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 provides a dynamic evolution expression recognition method, and the method comprises the steps: step 1, constructing and optimizing a past expression memory module according to the dynamic increase of expression types; step 2, according to the expression data of the new category, dynamically upgrading an expression recognition system by using a past expression memory module to minimize a'compact classification-distillation 'loss function; step 3, outputting an expression type label of the picture according to the input expression picture to be recognized and the expression recognition system in the current state. According to the invention, the facial expression recognition system can support training and recognition of any stage with dynamically increased categories, and stable and reliable performance is achieved. Compared with a traditional facial expression recognition system, the system significantly reduces memory and time loss during training.

Description

technical field [0001] The technical field involved in this application, in particular, relates to a method for recognizing dynamically evolving expressions and a system for recognizing dynamically evolving expressions. Background technique [0002] Expression recognition technology has been widely used in human-computer interaction, industrial control, medical education and other fields. It is an important tool for analyzing user behavior, evaluating psychological state, and monitoring disease development. With the continuous refinement of psychology for emotional analysis, more and more complex expression categories have been created, paid attention to and used by the public. In this case, the recognition method of dynamically evolving expressions becomes an urgent and fundamental challenge in expression recognition technology. [0003] However, in the existing technology, when using a fixed-type expression recognition method, when introducing data of a new expression cat...

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
IPC IPC(8): G06K9/00G06K9/62G06N3/04G06N3/08
CPCG06N3/04G06N3/08G06V40/176G06V40/168G06F18/2451
Inventor 赵曦滨朱俊杰骆炳君高跃
Owner TSINGHUA UNIV
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