Separation and recognition of overlapped latent images

Fingerprints are still the most important method for the recognition of criminal. Automatic Fingerprint Identification System (AFIS) has improved tremendously with storage, search and matching of fingerprints without human need. A fingerprint is an individual feature of all humans. No two fingers have identical ridge characteristics. In crime scenes, the latent images can be merged with some background images or more number of fingerprint images can be overlapped. Fingerprints obtained from crime places may be in low quality also. These kind of low quality, damaged and overlapped fingerprints create a tedious problem to identify and recognize the person. A critical step in AFIS is to automatically and reliably extract features from the fingerprint images. In this paper, we investigate the problem of retrieving the information of criminals using their images without human intervention. Our proposed work also describes the computational geometry to separate the overlapped images. The proposed work is planned to improve the performance of the system in terms of accuracy and speed with the increased number of samples for testing and training. This proposed research work tries to improve true acceptance rate, genuine acceptance rate and reduce false accept rate and false reject rate for latent images.