Implementation of perceptual resemblance of local plastic surgery facial images using Near Sets

When an individual undergo plastic surgery, the facial features are again constructed either in global or local way. Generally, this process changes the look of an individual. As these procedures become more and more common, future face recognition systems will be challenged to recognize individuals after plastic surgery has been performed. However, to alleviate the problems of face recognition after local plastic surgery of faces, we have proposed a system for perceptual resemblance of plastic surgery facial images using Near Sets. Here we have started with collection of Plastic surgery facial images database. In our work we have concentrated on three type of surgery Rhinoplasty, Blepharoplasty and Lip Augmentation. After obtaining both pre-surgery and post-surgery images we extract the features. We explore three different types of surgery with five facial features namely Average Grey, Normalized R, Normalized G, Normalized B, Shannon's Entropy. After feature extraction we found tolerance class of the features obtained. This tolerance class will tell us which object belongs to which class. Here Near set theory is used to find resemblance between objects in a different sets. The practical application of near set theory on the before and after plastic surgery facial images is to extract resemblance between them. Near set theory measure the degree of resemblance of facial images before and after plastic surgery. tHD, tNM and tHm is being used to measure the degree of similarity between plastic surgery images. tHD measure shows around 100% nearness as compared to tNM and tHM for all features. These measure can also be used in increasing the efficiency of any face recognition system.