Crime investigation and analysis using neural nets

Abstract

Crime investigation, with special reference to homicide cases, involves intensive and phase wise collection of data. We discuss a technique for comparing previously available data of the homicide cases with the current one. The system to be developed here is based upon an artificial neural network and is intended to be used by homicide detectives and other special task personnel for providing improved and sound reports to the courts of justice. The objective is thus to provide an automated crime analysis system (ACS) which is based upon an artificial neural network (ANN). The system also provides a crime visualization feature that gives the user a pictorial representation of crime details. This feature uses an incremental approach to represent the suspects and their concerns. Thus an elaborate visual description is achieved.