• Srushti Kantesh Kharat
• Sufiyan Patel

The proliferation of highly realistic deepfake technology poses a severe threat to the integrity of digital audio-visual evidence, challenging the legal and forensic communities. Current detection methods often operate on a single modality (audio or video) and struggle against the latest generative AI models, leading to unreliable authentication. Using a Design Science Research (DSR) methodology, this paper proposes a novel multi-modal forensic framework designed to authenticate audio-visual evidence by synergistically analyzing inherent spectral (audio) and spatial (video) signatures. Through a systematic literature review, this study identifies a critical gap in holistic, integrated deepfake forensic solutions. The core design artifact integrates a Convolutional Neural Network (CNN) for extracting spatial artifacts from video frames with a Recurrent Neural Network (RNN) for modeling temporal spectral features from audio waveforms. A fusion network then correlates these features to make a final authenticity determination, significantly enhancing robustness. A simulated evaluation on a diverse deepfake dataset demonstrates the framework's high efficacy, showing superior accuracy and generalizability compared to unimodal approaches, thereby providing a reliable tool for digital forensics.

"MULTI-MODAL FORENSIC FRAMEWORK FOR DEEPFAKE AUDIO-VISUAL EVIDENCE AUTHENTICATION USING SPECTRAL AND SPATIAL SIGNATURES", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 9, page no.e277-e280, September-2025, Available :http://www.jetir.org/papers/JETIR2509429.pdf

"MULTI-MODAL FORENSIC FRAMEWORK FOR DEEPFAKE AUDIO-VISUAL EVIDENCE AUTHENTICATION USING SPECTRAL AND SPATIAL SIGNATURES", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 9, page no. ppe277-e280, September-2025, Available at : http://www.jetir.org/papers/JETIR2509429.pdf