An Investigation of Crime Detection Using Artificial Intelligence and Face Sketch Synthesis

Abstract

Crime scene suspect prediction involves classifying potential perpetrators based on location, time, and crime type. Electronic forensics in large data scenarios poses challenges for investigators. Law enforcement relies on hand-drawn or computer-generated face sketches for perpetrator identification. To aid recognition, an automated method for face sketch expansion is essential. Deep learning models, like the Golden jackal optimized artificial neural network (GJO-ANN), generate face sketches for crime detection. Comparing these sketches with eyewitness and artist renditions helps identify similarities, enabling perpetrator detection. Experimental results validate the superior performance of GJO-ANN in face sketch synthesis for crime detection.