Joseph Chazalon

Joseph Chazalon is a Lecturer at EPITA (School of Engineering and Computer Science) in the LRDE research laboratory, specializing in image processing and pattern recognition. He teaches graduate courses in Machine Learning, Computer Vision, Deep Learning, GPU programming, and Docker for ML engineers.

His research develops novel methods for document analysis, including historical archive processing, hologram verification, and knowledge graph construction from archival sources. Current projects focus on weakly supervised learning for identity document verification and named entity recognition in structured historical documents.

Chazalon leads several collaborative projects involving historical map segmentation competitions (ICDAR 2021, 2024) and urban perpetual gazetteer development. His lab develops open benchmarks and datasets for reproducible research in document analysis.