# What is the right way to represent document images?

arXiv: https://arxiv.org/abs/1603.01076 · HTML: https://mlsanity.com/paper/1603.01076/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 3 Mar 2016.

## Built on by (most cited first, 5 of 5)

- [DocFormer: End-to-End Transformer for Document Understanding](https://mlsanity.com/paper/2106.11539/) (2021)
- [On Evaluation of Document Classification using RVL-CDIP](https://mlsanity.com/paper/2306.12550/) (2023)
- [EAML: Ensemble Self-Attention-based Mutual Learning Network for Document Image Classification](https://mlsanity.com/paper/2305.06923/) (2023)
- [Multimodal Side-Tuning for Document Classification](https://mlsanity.com/paper/2301.07502/) (2023)
- [DWT-CompCNN: Deep Image Classification Network for High Throughput JPEG 2000 Compressed Documents](https://mlsanity.com/paper/2306.01359/) (2023)

## Abstract (arXiv)

In this article we study the problem of document image representation based on visual features. We propose a comprehensive experimental study that compares three types of visual document image representations: (1) traditional so-called shallow features, such as the RunLength and the Fisher-Vector descriptors, (2) deep features based on Convolutional Neural Networks, and (3) features extracted from hybrid architectures that take inspiration from the two previous ones. We evaluate these features in several tasks (i.e. classification, clustering, and retrieval) and in different setups (e.g. domain transfer) using several public and in-house datasets. Our results show that deep features generally outperform other types of features when there is no domain shift and the new task is closely related to the one used to train the model. However, when a large domain or task shift is present, the Fisher-Vector shallow features generalize better and often obtain the best results.