Abstract
In this search, a new scheme based on facial imagery of suspect is used for automated criminal investigation. Because, classification of fingerprint by pixel-wise matching is tedious and the features based schemes often lead to misclassification and hence improper matching. The image matching algorithm attempts to partially match the facial image of the suspect with known images. The conventional model based approaches are difficult to be implemented. Unfortunately, with the increase in the complexity of the process being modeled, the difficulty in developing dependable fuzzy rules and membership functions increases. A novel approach based on Adaptive neuro-fuzzy is used. It has the benefits of both neural networks and fuzzy logic. The neuro-fuzzy hybrid system combines the advantages of fuzzy logic system, which deal with explicit knowledge that can be explained and understood, and neural networks, which deal with implicit knowledge, which can be acquired by learning. Fuzzy logic has tolerance for imprecision of data, while neural networks have tolerance for noisy data. The main trick in this matching lies in fuzzy membership, which keeps track of the important features in the human faces and their relative distances. The matching scheme has the advantages of size and rotational invariant. This means that the matching scheme is insensitive to variation of image size or their angular rotation on the facial image plane.
Keywords
criminal investigation
Facial Image
Fuzzy logic
Imprecision of Data
Member Ships
Neural Nets
Noisy Data
suspect