Facial Recognition for Crime Control: A Machine Learning-based Solution

The criminal record holds detailed information including criminal history and personal information with a photograph of an individual. In order to identify a criminal, eyewitness identification is necessary, despite often having poor image quality and unclear resolution. Identification can be performed through various means like DNA, fingerprint or eyes. Facial identification involves storing full and cropped images of criminals along with their personal and criminal information to create a database. Then another database is created to identify the criminal; eyewitnesses, with the help of a professional, try to piece together the criminal's face using saved cuts stored in a database. Then, using Amazon Recognition, we try to predict the criminal by comparing the created image with the existing database, if the result is 70-80%, we declare this face as a criminal. Amazon Recognition is an image recognition service that identifies objects, faces, etc. and find out the result from the database. This technology also has a feature to compare different faces, which we are using in our project.