I received the bachelor and the master degrees in computer science from the University of Monastir in 2015 and 2018, respectively. Then, I received the PhD degree from the Autonomous University of Barcelona in 2022. I am currently a post-doc researchr in the computer vision center (CVC). Previously, I was leading the machine learning team at Chordata Motion. Before that I was a pre-doc, then a post-doc researcher at CVC.
E-mail: msouibgui@cvc.uab.cat
I joined Chordata Motion to lead the machine learning team in developing intelligent motion capture/estimation systems.
I am participating in organizing the Privacy Preserving Federated Learning Document VQA (PFL-DocVQA) competition, which will be held in NeurIPS 2023. The goal of the competition is to develop privacy-preserving solutions for fine-tuning multi-modal language models for document understanding on distributed data.
I recived my PhD degree from the Autonomous University of Barcelona with the mark Excellent. From the committee composed of: Prof. Andreas Maier, Friedrich-Alexander Universität Erlangen-Nürnberg, Germany, Prof. Ernest Valveny, Universitat Autonoma de Barcelona, Spain and Dr. Naila Murray, Meta Artificial Intelligence Research (MAIR). United Kingdom
Our paper Text-DIAE: A Self-Supervised Degradation Invariant Autoencoders for Text Recognition and Document Enhancement was accepted in AAAI 2023.
The code, models and demo for our document image enhancement model based on transformer are now available on this repository
Working on explainable, secure and safe AI systems in the DocVQA field.
Developing intelligent motion capture systems.
My work is about developing privacy preserving machine learning models for document intelligence. I worked within the European Lighthouse on Secure and Safe AI (ELSA).
My work is about recognizing the handwritten historical ciphered manuscripts within the DECRYPT project.
Within the European project Elsa DECRYPT project, we develop safe and secure AI solutions. ELSA is a virtual center of excellence that will spearhead efforts in foundational safe and secure artificial intelligence (AI) methodology research. A large and growing network of top European experts in AI and machine learning is to promote the development and deployment of cutting-edge AI solutions in the future and make Europe the world's lighthouse of AI.
More Infos
Within the DECRYPT project, we release resources and tools with open access to facilitate research in historical cryptology, allowing collection, analysis and decryption of historical ciphertexts. Resources are collections of encrypted sources, and historical texts and language models. The tools facilitate the processing of the encrypted sources from transcription to decryption incl. cryptanalysis.
More InfosThis is a selected list. The full list of the publications is available on my Google Scholar Profile.
Souibgui, M. A., Fornés, A., Kessentini, Y., & Megyesi, B. (2022). Few Shots Are All You Need: A Progressive Few Shot Learning Approach for
Low Resource Handwritten text Recognition. Pattern Recognition Letters.
Paper on arxiv Code
Jemni, S. K.*, Souibgui, M. A.*, Kessentini, Y., & Fornés, A. (2021). Enhance to Read Better: A Multi-Task Adversarial Network for Handwritten Document Image Enhancement. Pattern Recognition.
Paper on arxiv
Souibgui, M. A., & Kessentini, Y. (2020). De-gan: A conditional generative adversarial network for document enhancement. IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI).
Paper on arxiv Code
Souibgui, M. A.*, Biswas, S.*, Mafla, A.*, Biten, A. F.*, Fornés, A., Kessentini, Y., Lladós, J, Gomez, L, \& Karatzas, D. (2023). Text-DIAE: A Self-supervised Degradation Invariant Autoencoders for Text Recognition and Document Enhancement. In AAAI Conference on Artificial Intelligence (AAAI).
Paper on arxiv Code
Souibgui, M. A.*, Biswas, S.*, Jemni, S. K.*, Kessentini, Y., Fornés, A., Lladós, J., & Pal, U. (2022). DocEnTr: An End-to-End Document Image Enhancement Transformer. In 2022 26th International Conference on Pattern Recognition (ICPR).
Paper on arxiv Code
Souibgui, M. A.*, Biten, A. F.*, Dey, S.*, Fornés, A., Kessentini, Y., Gomez, L., ... & Lladós, J. (2022). One-shot Compositional Data Generation for Low Resource Handwritten Text Recognition. In IEEE/CVF Winter Conference on Applications of Computer Vision (WACV).
Paper on arxiv
Chen, J., Souibgui, M. A., Fornés, A., & Megyesi, B. (2021). UnsupervisedAlphabet Matching in Historical Encrypted Manuscript Images. In 4th International Conference on Historical Cryptology (HistoCrypt) 2021.
Souibgui, M. A., Fornés, A., Kessentini, Y., & Tudor, C. (2020). A Few-shot Learning Approach for Historical Ciphered Manuscript Recognition. In 2020 25th International Conference on Pattern Recognition (ICPR) (pp. 5413-5420). IEEE.
Souibgui, M. A., Kessentini, Y., & Fornés, A. A Conditional GAN Based Approach for Distorted Camera Captured Documents Recovery. In 2020 4th Mediterranean Conference on Pattern Recognition and Artificial Intelligence (MedPRAI). Springer.
Chen, J., Souibgui, M. A., Fornés, A., & Megyesi, B. (2020, May). A Web-based Interactive Transcription Tool for Encrypted Manuscripts. In 3rd International Conference on Historical Cryptology (HistoCrypt) 2020 (No. 171, pp. 52-59). Linköping University Electronic Press.
Torras, P., Souibgui, M. A., Chen, J., & Fornés, A. (2021). A Transcription Is All You Need: Learning to Align through Attention. In International Workshop on Graphics Recognition (GREC).
* These authors contributed equally to the paper.
The awards recognise researchers from a CERCA center (Cata- lan Research Centers) who have just completed a doctoral thesis with clear market-oriented results. More information here !
Given by the ICPR2020 organizing committee for the paper entitled "A Few-shot Learning Approach for Historical Ciphered Manuscript Recognition in the track of Document and Media Analysis. More information here !