The aim of Crime Detection Through Facial Features Project is “Identification of a person in a crime scene by witness descriptions using Image Processing Techniques”.
Video fitting, for image reconstruction and person identification, is a well-known technique popularly used in scientific, industrial, medical and forensic applications. There has been a variety of “Application Software” based on Video fitting developed by a number of people.
Biometrics based personal identification provides a promising solution for this kind of problems. Biometrics features include face, speech, handwriting and finger prints etc. Face identification has the advantage of being a passive, non-invasive system for personal identification. The success of face identification depends on the solution of two problems: Representation and matching. Both the algebraic features and geometric features will be used to represent face images.
The work involved in Crime Detection Through Facial Features Project is:
- Obtaining the human face images from different video sources such as CCD cameras, Cam coders, Digital cameras, Scanners, stored images etc.
- These images are then digitized and stored in a memory for further processing.
- Then each video frame is segmented into different parts. For example, the given video frame containing a human face is sliced into three parts of all the images are then stored in a separate data bank. Similar parts of different images are stored in their respective files.
After creating the respective files of forehead, eyes and mouth, these files are used for reconstructing the images in order to get an identifiable image based on various descriptions.
The project titled FACE RECOGNITION USING IMAGE CROPPING is developed to design a system that supports to find the criminals in less time. A user with valid user_id and password are only allowed to access the developed system. In this project we store the details of the criminals in database and retrieve them when required.
Previously it was difficult to find the criminal manually through files. In the proposed system we can maintain the details of criminals of all regions in a database so that time required for finding the criminals is reduced.