Forensic text recovered from low resolution document thumbnails
Forensic Text Recovery from Document Thumbnails Using Text-Aware U-Net
Cryptography and Security
Summary
Documents deleted from a computer sometimes leave behind tiny, blurry thumbnail images that are hard to read. Standard tools that convert images to text struggle with these low-quality pictures. The authors created a new computer method that cleans up these thumbnails while keeping the letters and words true to their original shape. This method uses a special way to check if the recovered text matches the original, improving text recovery accuracy. Their approach helps recover readable words from very blurry thumbnails, making it easier to find hidden or deleted document contents.
What this means in practice
- •For digital forensics teams: Recover machine-readable text from deleted document thumbnails as evidence in legal investigations.
- •For document management services: Improve the readability of low-resolution thumbnails for better document indexing and retrieval.
Authors
Mohit Ajaykumar Dhabuwala, Weifeng Xu
Abstract
Forensic analysis of low-resolution Word thumbnails presents a significant challenge because these artifacts may constitute the only remaining visual evidence of deleted documents. Standard Optical Character Recognition (OCR) tools perform poorly on such severely degraded images, creating a critical bottleneck in forensic evidence recovery. Although deep learning architectures, such as U-Net, can improve visual image quality, models optimized solely using pixel-based loss functions tend to favor visual smoothness over the structural fidelity required for accurate text recognition. To address this limitation, we propose a text-aware restoration framework designed for forensic triage. The proposed approach employs a U-Net architecture guided by a hybrid optimization strategy that integrates pixel-level fidelity with machine-readable text constraints. Specifically, an OCR-based Normalized Levenshtein Distance (NLD) metric is used during model selection to prioritize the recovery of an accurate character topology. Experimental results demonstrate that the proposed method significantly reduces the NLD, enabling reliable recovery of machine-readable text from severely degraded thumbnails.