Bloodstain Classification in Forensic Analysis Using Optimized 3D CNN

Blood stain detection is essential in crime scene analysis as it provides valuable insights into the events that transpired, aids in identifying individuals involved and supports the investigation process of criminal cases. It helps investigators understand what happened, who was involved and when it occurred. Advancements in forensic science, particularly bloodstain analysis, have become imperative for enhancing crime scene reconstruction. Conventional methods like DNA analysis, chemical analysis often takes more time to identify bloodstain. Instead of DNA analysis this research contributes to the advancement of forensic science by introducing an innovative approach to bloodstain identification and classification by using a 3D CNN model utilizing the capabilities of Hyperspectral Imaging. Here we introduce an optimized 3D CNN with mish activation function and finding the best accuracy. This work will help to make faster investigations of forensic scene analysis and analyzing criminal cases. By applying the optimized 3D CNN model we got outperforms of the existing CNN model. We got 97% accuracy which is higher than the existing 3D CNN (95%) and hybrid CNN( 96%) model and parameters are reduced from 291895 to 125943.