Biometric Identification Based on PCA for Palmprint Feature Extraction

Biometric identification technologies, including palmprint recognition, have undergone rapid development in recent years. This paper uses a machine vision experiment platform to build a local palmprint recognition system that can collect initial palmprint images of multiple experimenters and classify them. Firstly, the MV-VS1200S machine vision platform can be used to obtain contactless palmprint images. Then, preprocess the initial palmprint images under different lighting conditions, by grayscale conversion, adaptive filtering and Butterworth high-pass filtering, to obtain palmprint images with clear texture features. Next, extract the region of interest (ROI) based on the center of gravity of the palm and the valley points between the fingers, and establish a template library for the ROI. Using the feature extraction algorithm based on principal component analysis, the palmprint recognition region was extracted, and the corresponding feature vector matrix was obtained. Finally, feature matching based on Euclidean distance was used to select the template palmprint image with the highest similarity for the measured palmprint, enabling identity recognition. The experimental results show that the overall recognition rate of the system can reach 85.4%.