Multimedia Forensics: Identification and Verification of Source Camera, Vehicle Speed Estimation, and Deepfakes Detection

This dissertation advances multimedia forensics by addressing three critical research areas that enhance the authenticity verification and analysis of digital media. Multimedia forensics, which encompasses techniques for examining images, videos, audio, and text, faces increasing challenges due to sophisticated editing tools and massive data volumes. In the first study, a fast source camera identification and verification method based on PRNU analysis is proposed for video forensic investigations. By integrating camera rolling and I-frame analysis, this approach achieves a processing speed improvement of at least 15 times over conventional frame-by-frame methods while reducing false positives. The second study focuses on vehicular speed estimation from dashboard camera footage. By applying photogrammetry and cross-ratio techniques to dynamically extracted spatial-temporal data, the method accurately estimates vehicle speed with minimal reliance on lane markings—yielding deviations below 1 km/h in single-vehicle scenarios and under 3 km/h when multiple vehicles are present. The third study addresses the detection of deepfake media. A detection framework is developed that leverages no-reference image quality assessment (BRISQUE, NIQE, and PIQUE) combined with support vector machine classification. Evaluations of benchmark datasets demonstrate robust performance under degraded conditions, achieving up to 98% accuracy. Additionally, an integrated method is introduced that further improves detection accuracy for cross-dataset evaluation by up to 7%. Collectively, these contributions enhance the efficiency, accuracy, and robustness of forensic analysis, thereby bolstering the credibility and trustworthiness of digital media across various applications.