Font Size: a A A

The Frame Design Of The Meeting Record System And The Realization Of Face Recognition Technology In The System

Posted on:2020-03-12Degree:MasterType:Thesis
Country:ChinaCandidate:Y C YangFull Text:PDF
GTID:2428330578982360Subject:Software engineering
Abstract/Summary:PDF Full Text Request
As we all know,meetings are an important link in daily work,such as information transmission,atmosphere building and cost saving.Therefore,how to improve the efficiency of meeting minutes becomes particularly important.With the rapid development of face recognition technology face recognition technology has been more and more applied in various fields of life.The application of face recognition technology in meeting record system can not only solve the efficiency problem in daily meeting record to a certain extent,but also share the pressure of meeting record personnel,which has important practical application value.At the same time,face recognition technology has natural intuitive,optional,a non-contact,concurrently,and many other advantages,these advantages and meeting minutes need to identify the speaker identity,whether people speak in the current state of the demand is very fit,so face recognition technology was applied to meeting record system development and implementation,can give the meeting record of the existing system to a good supplement.In the actual development,the open source of face recognition technology and good selfcorrection are also the key factors for the better presentation of the meeting record system.Based on this,this paper will optimize and improve the efficiency of meeting minutes as the core demand,put forward a meeting minutes system solution based on face recognition technology.The whole meeting recording system mainly consists of three parts: speech recognition,face recognition and meeting recording system framework construction.This paper focuses on the face recognition and meeting recording system framework construction.In the face recognition part,through the reasonable utilization of multi-task cascade model and “landmask” face feature point extraction,and combined with the innovative speaker recognition algorithm in this paper,the two requirements of identifying the identity of participants and judging the speaker at the current moment were fulfilled.The open source Django framework is used to complete the design and implementation of background management module,foreground module,data recording module,identity recognition module,and whether to speak or not.In the end,through functional test and performance test,the reasonable application of face recognition technology can be verified to solve the development and implementation of the meeting record system.
Keywords/Search Tags:face recognition, meeting record system, multi-task cascade, landmask, dlib, speaker recognition
PDF Full Text Request
Related items