Font Size: a A A

Research Of Classification And Recognition Technology Based On The Infrared Thermal Image Of Handprint

Posted on:2019-03-29Degree:MasterType:Thesis
Country:ChinaCandidate:S L YuFull Text:PDF
GTID:2428330551961932Subject:Computer technology
Abstract/Summary:PDF Full Text Request
In the information era,biometric feature identification technology has become one of the most significant measures for personal authentication.Nowadays,visible images are always treated as the candidate objects for vast major of identification methods,which potentially increases the difficulty for extracting the hidden biometric features in some scenarios.To solve the bottlenecks mentioned above,the topic focuses on taking the advantage of infrared thermal imaging in identification so that the robustness of the identification system can be potentially improved in some scenarios.We conduct a detailed study around identification and classification of the handprint infrared thermal image,go into depth on the problems encountered in image acquisition,image preprocessing,feature extraction and algorithm recognition.In terms of image acquisition,the paper uses infrared camera to get multiple handprint infrared thermal images as the real-world experimental dataset.After that,the paper studies the methods of pretreating handprint infrared thermal images.The contour of the handprint target can be obtained by processing the image gray-scale,noise reduction and contour extraction.To excavate the latent biometric information contained in the handprint infrared thermal image,this paper employ two curve fitting methods for representing the feature of handprint's shape.Then,propose a novel feature extraction method based on multi-region elliptic fitting.In the classification phase,the paper implements two classification methods,i.e.k-nearest neighbor and support vector machine.Through adjusting the distance calculation formula and increasing the dimension of feature vectors,we finally obtain different recognition effects of the handprint infrared thermal image in a variety of situations.The accuracy rate reaches up to 92.00%.The experimental results proved that it is feasible to authenticate the personal identification by using the infrared thermal image of handprint,and the method can be used as auxiliary identification method in forensic identification,investigation,etc.
Keywords/Search Tags:biometric recognition, handprint thermal image, ellipse fitting, feature extraction
PDF Full Text Request
Related items