Digital Forensic Based Object Recognition for Enhanced Crime Scene Interpretation
Principal Software Engineer, IT, Societe Generale, India, Vikash Vikash, Technical Lead, IT, Siemens Healthineers, India, Durga Sivashankar, Siddharth Sriram, Manish Nagpal, Warish D. Patel, Shweta Loonkar · Journal of Intelligent Systems and Internet of Things · 2024
This research introduces a novel and comprehensive framework for digital forensics-based crime scene interpretation. The proposed framework comprises five algorithms, each serving a distinct purpose in enhancing image quality, extracting features, matching, and constructing a database, recognizing, and reconstructing objects in 3D, and conducting context-aware analysis. An ablation study validates the necessity of each algorithmic step. The framework consistently outperforms existing methods in terms of accuracy, precision, recall, and processing time. A detailed comparative analysis of parameters further highlights its cost-effectiveness, moderate complexity, superior data integration, and scalability. Visualizations underscore its dominance across multiple metrics and parameters, positioning it as an advanced solution for digital forensic-based object recognition in crime scene interpretation.