FINGERPRINT ATTENDENCE SYSTEM WITH LIVENESS DETECTION

Over the years, the manual method of taking attendance in any higher institution of learning are often violated because they can easily manipulate the manual attendance register. The attendance of the students is essential for the progress of an institution as they help to move the institution forward. Due to the uniqueness and consistency of fingerprints, they have been used for identification over many years and is being automated due to advancement in computing capabilities. Taking the advantage of recent technologies, we have developed an automatic fingerprint-based attendance registering system using Slim-Residual Convolutional Neural Network (Modified-CNN) and Octantal Neatest-Neighbourhood Structure (ONNS). Firstly, the CNN score feature vector is processed to discriminate between live fingerprints and fake ones. Secondly, the ONNS method is used to find closest minutia, from each sector of octant, to the central minutia (core point of the fingerprint) and used for fingerprint matching with the fingerprints of the students maintained in the database. Experimental results are carried on the images collected from the students to show the results of the proposed algorithms.