WordVIS: A Color Worth A Thousand Words

TL;DR

This paper introduces WordVIS, a method that embeds textual features into visual space, enabling lightweight image classifiers to perform well on document classification with small datasets, reducing data and computational needs.

Contribution

The paper presents a novel approach to embed textual features into visual space, improving image-based document classification accuracy on limited data without extensive training.

Findings

Achieved 4.64% improvement with ResNet50 without pre-training.

Set a new record of 91.14% accuracy on Tobacco-3482 dataset.

Demonstrated effectiveness of lightweight classifiers with embedded textual features.

Abstract

Document classification is considered a critical element in automated document processing systems. In recent years multi-modal approaches have become increasingly popular for document classification. Despite their improvements, these approaches are underutilized in the industry due to their requirement for a tremendous volume of training data and extensive computational power. In this paper, we attempt to address these issues by embedding textual features directly into the visual space, allowing lightweight image-based classifiers to achieve state-of-the-art results using small-scale datasets in document classification. To evaluate the efficacy of the visual features generated from our approach on limited data, we tested on the standard dataset Tobacco-3482. Our experiments show a tremendous improvement in image-based classifiers, achieving an improvement of 4.64% using ResNet50 with no document pre-training. It also sets a new record for the best accuracy of the Tobacco-3482 dataset with a score of 91.14% using the image-based DocXClassifier with no document pre-training. The simplicity of the approach, its resource requirements, and subsequent results provide a good prospect for its use in industrial use cases.