Font Size: a A A

Designand Transplantation For Recognition Algorithms Of Palm Vein Images

Posted on:2016-06-24Degree:MasterType:Thesis
Country:ChinaCandidate:S L LvFull Text:PDF
GTID:2308330482467970Subject:Control Engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of information technology, especially the Internet technology, personal information security has attracted a lot of attention. Compared with the traditional personal authentication, palm vein recognition method is much more reliable, and it has become one of the hotspots in the current information technology.Analyzing the methods for the preprocessing and the recognition of the palm vein images, this thesis designs the algorithms, and migrates a part of preprocessing algorithms to FPGA chip. The contents and the achievements of this thesis are listed as follows:(1)Design of preprocessing algorithm for palm vein images: This thesis adopts extracting of the ROI(Region Of Interest), normalization, smoothing and enhancing processing for the images. Firstly, we propose a global threshold method to apply binaryzation for the original palm vein image, then extract the ROI, rotate and normalize it, smooth the image using anisotropic diffusion. Finally, a logarithmic gray-scale transformation method is used to enhance the image.(2)Design of feature extraction and recognition algorithm: This thesis adopts the matching algorithm based on feature points, using local feature vectors and Euclidean distance for feature matching. In registration mode, 5 images are randomly selected from 15 images of each palm vein as the training samples, then the rest of the images as verification samples, and get the matching template by training. In Identification mode, each test vein image is matched with all matching templates in the database, and get the correct recognition rate.(3)The image algorithm and software are migrated to the palm vein acquiring hardware platform, including Linux operating system and Opencv Library. The preprocessed image is transmitted to the PC(Personal Computer) through the serial port, and the final recognition is achieved in the PC.(4)Finally, summarizing the full text content, and proposing improvement and prospect to the running time of algorithm and the degree of algorithm transplantation.
Keywords/Search Tags:palm vein, biological recognition, image preprocessing, feature extraction
PDF Full Text Request
Related items