A Prototype Design and Image Quality Assessment of Low-Cost Finger Vein Image Acquisition using Logitech Webcam
Keywords:Finger Vein, Image Processing, Graphical User Interface (GUI), CCD, Near-infrared LED
Finger vein pattern of each person varies one after another because of the blood vessels patterns are unique to everyone. Also, blood vessels are hidden beneath the skin's surface, and thus, it is nearly impossible to imitate them. Hence, this study proposed a finger veins image acquisition system by using a Logitech Webcam. This Logitech Webcam is a Charge-Coupled Device (CCD) image sensor that can see the finger veins with naked eyes. The lights from double-arrangement of NIR LEDs (940 nm) penetrated the finger and the CCD image sensor captured the finger images. The double line arrangement was chosen for the prototype as the finger images taken using this method has the lowest Mean Square Error (MSE) (186.04) and the highest Peak Signal-to-Noise Ratio (PSNR) (25.47) values. In total, 15 subjects were recruited to test the functionality of the developed system. The image of finger for each subject was taken five times and they were analysed using MSE and PSNR to determine the best quality of image. The best image with the highest PSNR from each subject was then used to extract the finger vein using edge segmentation (edge detection method) and feature extraction. Sobel and Canny were used for edge detection whereas Wide Line Detector, Maximum Curvature Method and Repeated Line Tracking were used for feature extraction. For the image segmentation, edge detection was used to detect the borderlines of the finger vein images, making it easier to see the differences between each technique under edge detection. In feature extraction, repeated line tracking method presents the best feature extraction as it can show the remarkable lines of the vein, and this will be beneficial for the biometric application. Finally, the developed graphical user interface (GUI) was able to show how the captured image of finger undergo image pre-processing (grayscale converter, image enhancement, image cropping and image resize), edge segmentation and feature extraction methods.