Tiny CNN for feature point description for document analysis: approach and dataset

TL;DR

This paper introduces a lightweight neural network for feature point description in document analysis, supported by a new dataset and training method, demonstrating effectiveness on document and patch matching tasks with limited computational resources.

Contribution

It presents a new dataset and training approach specifically for lightweight neural networks in document feature point description, addressing resource constraints.

Findings

The proposed network performs well on document matching tasks.

The new dataset improves training effectiveness for lightweight models.

The approach outperforms existing methods on complex background datasets.

Abstract

In this paper, we study the problem of feature points description in the context of document analysis and template matching. Our study shows that the specific training data is required for the task especially if we are to train a lightweight neural network that will be usable on devices with limited computational resources. In this paper, we construct and provide a dataset with a method of training patches retrieval. We prove the effectiveness of this data by training a lightweight neural network and show how it performs in both documents and general patches matching. The training was done on the provided dataset in comparison with HPatches training dataset and for the testing we use HPatches testing framework and two publicly available datasets with various documents pictured on complex backgrounds: MIDV-500 and MIDV-2019.