This thesis is in line with the domain of handwriting recognition and document analysis. The goal is to design an end-to-end model able to understand the inherent structure of documents, combined with text recognition. We focus on historical newspapers for the valuable data they contain. We are interested in designing an advanced…
PhD Student: Swan Serre Supervisors: Eric Anquetil, Nathalie Girard
The aim of this thesis will be to look at self-supervised learning in order to take advantage of unsu-pervised examples for handwriting recognition. In recent years, a great deal of work has focused onself-supervised approaches for network pre-training. Starting with a pretext task that does not requiremanual annotation, the model…
Shadoc team works with the ANTAI national agency on text recognition by reducingsyntactic bias modeling and integrating knowledge into models. This work is appliedto license plates and handwritten documents.