ENHANCED FINGER VEIN VERIFICATION WITH DEEP REPRESENTATION-BASED FEATURE EXTRACTION AND RECOVERY USING LOCAL LINE BINARY PATTERN DESCRIPTOR

This approach presents a finger vein personal verification mechanism. Due to the attributes are internal to the human body, finger veins are thought to be more secure than other biometric traits that are based on the hands, like fingerprints and palm prints. A novel texture descriptor known as the local line binary pattern (LLBP) is used as a feature extraction tool in the suggested method. Local binary pattern (LBP) has a square neighbourhood shape, whereas LBP's neighbourhood form is a straight line. The suggested approach employing LBP performs better than the prior methods using LBP, according to experimental findings. With little a priori knowledge, the deep learning model proposed in this research can extract and retrieve vein characteristics. First, using a combination of the most cutting-edge handmade finger-vein picture segmentation methods, we instantly recognise finger-vein patterns and backdrop may be distinguished into two distinct regions: a clear zone as well as an ambiguous region with little separability across them. The first refers to pixels that are assigned the same segmentation label (foreground or background) by all of the aforementioned segmentation algorithms, while the second describes all other pixels. This method is used to categorise the pixels in the clear zone as foreground or background and automatically eliminate the ambiguous region.