Image processing based Latent Fingerprint Forensics - A Survey

Latent fingerprints are the prints not visible to the naked eye and are impressions left at crime scenes unintentionally. Feature extraction is a substantial step towards latent fingerprints forensics. The factors affecting the image segmentation and quality enhancement of latent fingerprints can be the variation in the quality of image, continuous noise barrier, fingerprint lifting mechanism, overlapping of fingerprints. The intricacies of correct decision-making, policies impact the manual effort to be reduced to an automated system. The Frame of reference is challenged in pre-feature extraction and post-feature extraction phases of latent fingerprint analysis leading to erroneous decision making, influencing image segmentation. An intelligent automated system runs on controlling and learning the variation in quantitative and qualitative factors. Whereas segmentation perceives the quality of decision making as secondary to the need to address the parameters affecting the process, use of efficient and reliable estimation is considered a sufficient argument to impetus automation and thereby giving freedom to the rate of error in the output.