Abstract: In this thesis, we propose different contributions with the goal of enhancing and recognizing historical handwritten document images, especially the ones with rare scripts, such as cipher documents.
In the first part, some effective end-to-end models for Document Image Enhancement (DIE) using deep learning models were presented. First, Generative Adversarial Networks (cGAN) for different tasks (document clean-up, binarization, deblurring, and watermark removal) were explored. Next, we further improve the results by recovering the degraded document images into a clean and readable form by integrating a text recognizer into the cGAN model to promote the generated document image to be more readable. Afterward, we present a new encoder-decoder architecture based on vision transformers to enhance both machine-printed and handwritten document images, in an end-to-end fashion.
The second part of the thesis addresses Handwritten Text Recognition (HTR) in low resource scenarios, i.e. when only few labeled training data is available. We propose novel methods for recognizing ciphers with rare scripts. First, a few-shot object detection based method was proposed. Then, we incorporate a progressive learning strategy that automatically assignspseudo-labels to a set of unlabeled data to reduce the human labor of annotating few pages while maintaining the good performance of the model. Secondly, a data generation technique based on Bayesian Program Learning (BPL) is proposed to overcome the lack of data in such rare scripts. Thirdly, we propose a Text-Degradation Invariant Auto Encoder (Text-DIAE). This latter self-supervised model is designed to tackle two tasks, text recognition and document image enhancement. The proposed model does not exhibit limitations of previous state-of-the-art methods based on contrastive losses, while at the same time, it requires substantially fewer data samples to converge.
In the third part of the thesis, we analyze, from the user perspective, the usage of HTR systems in low resource scenarios. This contrasts with the usual research on HTR, which often focuses on technical aspects only and rarely devotes efforts on implementing software tools for scholars in Humanities.

Abstract: This paper introduces the DocILE benchmark with the largest dataset of business documents for the tasks of Key Information Localization and Extraction and Line Item Recognition. It contains 6.7k annotated business documents, 100k synthetically generated documents, and nearly 1M unlabeled documents for unsupervised pre-training. The dataset has been built with knowledge of domain- and task-specific aspects, resulting in the following key features: (i) annotations in 55 classes, which surpasses the granularity of previously published key information extraction datasets by a large margin; (ii) Line Item Recognition represents a highly practical information extraction task, where key information has to be assigned to items in a table; (iii) documents come from numerous layouts and the test set includes zero- and few-shot cases as well as layouts commonly seen in the training set. The benchmark comes with several baselines, including RoBERTa, LayoutLMv3 and DETR-based Table Transformer; applied to both tasks of the DocILE benchmark, with results shared in this paper, offering a quick starting point for future work. The dataset, baselines and supplementary material are available at https://github.com/rossumai/docile.

Abstract: Transcription of handwritten text in (old) documents is an important, time-consuming task for digital libraries. It might be carried out by first processing all document images off-line, and then manually supervising system transcriptions to edit incorrect parts. However, current techniques for automatic page layout analysis, text line detection and handwriting recognition are still far from perfect, and thus post-editing system output is not clearly better than simply ignoring it.
A more effective approach to transcribe old text documents is to follow an interactive- predictive paradigm in which both, the system is guided by the user, and the user is assisted by the system to complete the transcription task as efficiently as possible. Following this approach, a system prototype called GIDOC (Gimp-based Interactive transcription of old text DOCuments) has been developed to provide user-friendly, integrated support for interactive-predictive layout analysis, line detection and handwriting transcription.
GIDOC is designed to work with (large) collections of homogeneous documents, that is, of similar structure and writing styles. They are annotated sequentially, by (par- tially) supervising hypotheses drawn from statistical models that are constantly updated with an increasing number of available annotated documents. And this is done at different annotation levels. For instance, at the level of page layout analysis, GIDOC uses a novel text block detection method in which conventional, memoryless techniques are improved with a “history” model of text block positions. Similarly, at the level of text line image transcription, GIDOC includes a handwriting recognizer which is steadily improved with a growing number of (partially) supervised transcriptions.

Abstract: According to the Cambridge Dictionaries Online, a symbol is a sign, shape, or object that is used to represent something else. Symbol recognition is a subfield of general pattern recognition problems that focuses on identifying, detecting, and recognizing symbols in technical drawings, maps, or miscellaneous documents such as logos and musical scores. This chapter aims at providing the reader an overview of the different existing ways of describing and recognizing symbols and how the field has evolved to attain a certain degree of maturity.