Enhanced Image Reconstruction and Contextual Intelligence For Cyber-Forensics

The low-resolution image combined with noise and blurriness makes it difficult for investigators to use the data in legal proceedings. When combined with noise and blurriness these interventions make forensic evidence unusable for legal proceedings. Real ESRGAN used in the proposed automated system produces image enhancement results with PSNR at 30.2 and SSIM at 0.88. In a proposed model the images reach 1024×1024 resolution post restoration for enhanced quality that results in SSIM 0.92 and PSNR 32.1 which improves text and facial analytics. The performance tests demonstrate how the OCR module achieves 95.6% accuracy rate in extracting text from enhanced noisy images. The forensic analysis benefits from Gemini LLM through its scene description capabilities that enhance crime scene and document interpretation. The 50 images per minute processing speed of this technology supports both criminal investigation needs and historical conservation efforts alongside intelligence gathering work.