PatFig: Generating Short and Long Captions for Patent Figures

TL;DR

This paper presents Qatent PatFig, a large-scale dataset of patent figures with captions and annotations, enabling the development and evaluation of models for automatic patent figure captioning.

Contribution

The paper introduces a new extensive dataset for patent figure captioning and demonstrates its utility by fine-tuning a model to generate descriptive captions.

Findings

Fine-tuned LVLM model effectively generates patent figure descriptions.

Incorporating text cues improves captioning accuracy.

Dataset supports future research in patent image understanding.

Abstract

This paper introduces Qatent PatFig, a novel large-scale patent figure dataset comprising 30,000+ patent figures from over 11,000 European patent applications. For each figure, this dataset provides short and long captions, reference numerals, their corresponding terms, and the minimal claim set that describes the interactions between the components of the image. To assess the usability of the dataset, we finetune an LVLM model on Qatent PatFig to generate short and long descriptions, and we investigate the effects of incorporating various text-based cues at the prediction stage of the patent figure captioning process.