Leveraging Computer Vision in AI for Efficient Image Analysis in Case Investigations

This paper presents a novel approach using computer vision algorithms for image analysis facilitating regulators' investigation. Digital evidence is rapidly increasing, originating from selected data viewed through forensic tools. While conventional tools use keywords and metadata search, they fall short of the expectation when it comes to multimedia images on phones. Visual inspection is still relied upon for image selection. Manual selection becomes limited in complex cases with increasing image volumes. The proposed method addresses these challenges through Object Detection, Face Detection, and Optical Character Recognition (OCR). Object Detection classifies evidence, Face Detection identifies associates and OCR extracts text for keyword search and personal information identification. This i2i (images to insights) method enables fast and relevant image selection, aiding prosecution and ensuring search scope adherence. It also assists in redaction and handling privileged claims in its extended use.