Transparency Techniques for Neural Networks trained on Writer Identification and Writer Verification

TL;DR

This paper explores the application of two transparency techniques to neural networks in writer identification and verification, demonstrating that pixel-wise saliency maps effectively support forensic experts by highlighting relevant handwriting features.

Contribution

First application of transparency techniques to neural networks in writer identification and verification, with evaluation showing pixel-wise saliency maps outperform point-specific maps.

Findings

Pixel-wise saliency maps outperform point-specific maps.

Saliency maps are useful for forensic support.

Transparency techniques align with expert considerations.

Abstract

Neural Networks are the state of the art for many tasks in the computer vision domain, including Writer Identification (WI) and Writer Verification (WV). The transparency of these "black box" systems is important for improvements of performance and reliability. For this work, two transparency techniques are applied to neural networks trained on WI and WV for the first time in this domain. The first technique provides pixel-level saliency maps, while the point-specific saliency maps of the second technique provide information on similarities between two images. The transparency techniques are evaluated using deletion and insertion score metrics. The goal is to support forensic experts with information on similarities in handwritten text and to explore the characteristics selected by a neural network for the identification process. For the qualitative evaluation, the highlights of the maps are compared to the areas forensic experts consider during the identification process. The evaluation results show that the pixel-wise saliency maps outperform the point-specific saliency maps and are suitable for the support of forensic experts.