Hybrid AI Framework for Bloodstain Detection, Classification and Forensic Scene Reconstruction

Abstract

Bloodstain Pattern Analysis is used in forensic investigations for reconstructing the sequence of events that occurred at a crime scene. To improve the forensic reasoning, we use Retrieval augmented Generation that is integrated in the pro- posed system. To classify the bloodstain into the given categories we use YOLO based detection model. The bloodstain is further analyzed by Morphological and spatial features and the results are sent to RAG engine. This RAG engine uses Forensic literature and do aim specific resources and historic case to reconstruct the crime scene. Feedback from the expert forensic officers is used to reinforce and increase the accuracy of the reasoning engine. The framework connects AI driven pattern recognition with expert forensic knowledge, resulting in explainable and reliable tool which assists investigators in decision making and evidence interpretation faster.