Hand Recognition using Palm and Hand Geometry Features
Irfan Ahmad, Zahoor Jan, Inayat Ali Shah, Jamil Ahmad · 2017
Biometric recognition is an emerging technology and attaining high performance from last several years. Our proposed system focuses on multiple features derived from a single template. The aim of this work is to combine palm print and hand geometry feature to achieve accuracy and high performance. The features of interest in proposed system are hand geometry, palm length, palm width and palm ratio. Hand geometry consists of hand length, hand width, fingers length and fingers width. These features are integrated to form a combined feature vector. Proposed system extracts only those features which are invariant to small variations in position of hand. A total of 20 features are obtained from each image by creating bounding box of each object, a combined feature vector of extracted 20 features are formed. An average of combined feature vector is calculated by comparing the combined feature vector with stored templates. The proposed system efficiency is calculated using true positive, true negative, false positive, false negative, accuracy and runtime. The system achieves an accuracy of 95.5%.