Improving and Evaluating Machine Learning Methods for Forensic Shoeprint Matching

arXiv:2405.14878v1 Announce Type: cross
Abstract: We propose a machine learning pipeline for forensic shoeprint pattern matching that improves on the accuracy and generalisability of existing methods. We extract 2D coordinates from shoeprint scans using edge detection and align the two shoeprints with iterative closest point (ICP). We then extract similarity metrics to quantify how well the two prints match and use these metrics to train a random forest that generates a probabilistic measurement of how likely two prints are to …

abstract accuracy arxiv cs.cv cs.lg detection edge eess.iv extract improving iterative machine machine learning pattern pipeline scans stat.ap type