Block Statistic for Palmprint Recognition Based on High Frequency Coefficients under Wavelet Transform

2012 ◽  
Vol 182-183 ◽  
pp. 1287-1291
Author(s):  
Yu Qin Liu ◽  
Wei Qi Yuan ◽  
Jin Yu Guo
Keyword(s):  
Wavelet Transform ◽  
High Frequency ◽  
Nearest Neighbor ◽  
Feature Vector ◽  
Region Of Interest ◽  
Palmprint Recognition ◽  
Nearest Neighbor Classifier ◽  
Palmprint Identification ◽  
The Mean ◽  
Neighbor Classifier

Palmprint recognition for identification provides a new scheme for information security. This paper presents a block statistic method based on high frequency coefficients under wavelet transform for palmprint identification. Firstly, the method decomposes region of interest (ROI) of the palmprint with the wavelet transform. Then it blocks the high-frequency sub-image. The mean and the variance for each sub-block are found. All the means and the variances constitute feature vector for the image. At last the nearest neighbor classifier is used to classify the images. The method was tested on the basis of UST palmprint image database. From the experimental results, the method can satisfy the uses without excessive demands for collection images.

scholarly journals An Efficient Method for Similarity Measure in Independent PCA based Classification

2013 ◽  
Vol 9 (3) ◽  
pp. 1099-1109
Author(s):  
Dr. H. B. Kekre ◽  
Dr. Tanuja K. Sarode ◽  
Jagruti K. Save
Keyword(s):  
Coordinate System ◽  
Similarity Measure ◽  
Efficient Method ◽  
Nearest Neighbor ◽  
Feature Vector ◽  
Test Image ◽  
Nearest Neighbor Classifier ◽  
Feature Vectors ◽  
The Given ◽  
Neighbor Classifier

The paper presents a new approach of finding nearest neighbor in image classification algorithm by proposing efficient method for similarity measure. Generally in supervised classification, after finding the feature vectors of training images and testing images, nearest neighbor classifier does the classification job. This classifier uses different distance measures such as Euclidean distance, Manhattan distance etc. to find the nearest training feature vector. This paper proposes to use Mean Squared Error (MSE) to find the nearness between two images. Initially Independent Principal Component Analysis (PCA),which we discussed in our earlier work, is applied to images of each class to generate Eigen coordinate system for that class. Then for the given test image, a set of feature vectors is generated. New images are reconstructed using each Eigen coordinate system and the corresponding test feature vector. Lowest MSE between the given test image and new reconstructed image indicates the corresponding class for that image. The experiments are conducted on COIL-100 database. The performance is also compared with  distance based nearest neighbor classifier. Results show that the proposed method achieves high accuracy even for small size of training set.

A Palmprint Recognition Approach Based on Image Segmentation of Region of Interest

2016 ◽  
Vol 30 (02) ◽  
pp. 1656002 ◽  
Author(s):  
Qing E Wu ◽  
Zhiwu Chen ◽  
Ruijie Han ◽  
Cunxiang Yang ◽  
Yuhao Du ◽  
...  
Keyword(s):  
Image Segmentation ◽  
Wavelet Transform ◽  
Recognition Rate ◽  
Region Of Interest ◽  
Palmprint Recognition ◽  
Characteristic Points ◽  
Positioning Method ◽  
Roi Extraction ◽  
Energy Feature

To carry out an effective recognition for palmprint, this paper presents an algorithm of image segmentation of region of interest (ROI), extracts the ROI of a palmprint image and studies the composing features of palmprint. This paper constructs a coordinate by making use of characteristic points in the palm geometric contour, improves the algorithm of ROI extraction and provides a positioning method of ROI. Moreover, this paper uses the wavelet transform to divide up ROI, extracts the energy feature of wavelet, gives an approach of matching and recognition to improve the correctness and efficiency of existing main recognition approaches, and compares it with existing main approaches of palmprint recognition by experiments. The experiment results show that the approach in this paper has the better recognition effect, the faster matching speed, and the higher recognition rate which is improved averagely by 2.69% than those of the main recognition approaches.

Sign in / Sign up

Export Citation Format

Share Document