Fixed-sized representation learning from Offline Handwritten Signatures of different sizes

TL;DR

This paper introduces a method using Spatial Pyramid Pooling to learn fixed-sized feature representations from variable-sized offline handwritten signatures, improving flexibility and performance in signature verification tasks.

Contribution

It proposes a novel neural network architecture modification that removes size constraints, allowing effective signature verification across different sizes and resolutions, with demonstrated improvements.

Findings

Higher resolutions (300-600dpi) improve performance with skilled forgeries.

Lower resolutions (~100dpi) suffice when only genuine signatures are used.

Fine-tuning enhances performance under different acquisition conditions.

Abstract

Methods for learning feature representations for Offline Handwritten Signature Verification have been successfully proposed in recent literature, using Deep Convolutional Neural Networks to learn representations from signature pixels. Such methods reported large performance improvements compared to handcrafted feature extractors. However, they also introduced an important constraint: the inputs to the neural networks must have a fixed size, while signatures vary significantly in size between different users. In this paper we propose addressing this issue by learning a fixed-sized representation from variable-sized signatures by modifying the network architecture, using Spatial Pyramid Pooling. We also investigate the impact of the resolution of the images used for training, and the impact of adapting (fine-tuning) the representations to new operating conditions (different acquisition protocols, such as writing instruments and scan resolution). On the GPDS dataset, we achieve results comparable with the state-of-the-art, while removing the constraint of having a maximum size for the signatures to be processed. We also show that using higher resolutions (300 or 600dpi) can improve performance when skilled forgeries from a subset of users are available for feature learning, but lower resolutions (around 100dpi) can be used if only genuine signatures are used. Lastly, we show that fine-tuning can improve performance when the operating conditions change.