ABSTRACT
The facial structure is unique to each person, and it can be identified by extracting various factors that depict them. Therefore, it is well suited for human identification (ID). In this project, we propose the concept of using facial recognition as a means of human identification for increasing security of bank lockers. For the locker system of banks in India the only method of accessing the locker is through a key that is first provided at the time of purchasing the locker space. Here we wish to increase the security of accessing the locker. When a person purchases a locker along with supplying the locker key, the person’s facial image is acquired using a camera (normal webcam). The face is detected using Viola jones algorithm and facial features i.e Histogram of oriented Gradients are extracted. These HOG coefficients are extracted for all customers and a database is created. The next time the person arrives to access his locker his facial features are again scanned and the obtained values are compared against the previously recorded values in the database to check whether the customer is genuine or not. The person is authenticated when the extracted features and the features stored in the database show the same or nearly similar numerical values. Once the customer has been authenticated by the system, a confirmation text message is sent to the respective person’s phone to notify him that he is permitted to access the locker. The experimental results show that facial recognition increases the security of the bank locker system.
