Logo Lab-STICCDes capteurs à la connaissance : Communiquer et décider

TOUATI Mohamed

Post doctorant - Université de Bretagne Occidentale

263123841_5065028483530611_6865831888663362935_n.jpeg
Équipe : IRIS  
Fonction : ATER
Téléphone : 0601442600
Adresse email : 
Adresse : UBO Site de Brest - LC 115A - 2 rue françois verny - 29200 BREST


As a PhD Researcher at Lab-STICC (UMR CNRS 6285), my work primarily focuses on medical imaging and the classification of diabetic retinopathy. I am engaged in cutting-edge research aimed at leveraging artificial intelligence and deep learning technologies to enhance the detection and diagnosis of this disease. My expertise spans various aspects of Convolutional Neural Networks (CNNs), Transformers, and advanced image processing techniques, with a particular emphasis on data balancing and improving model accuracy.

I aim to improve medical diagnosis through automation and highly accurate models, while addressing challenges such as data imbalance and information security in medical images.

My Expertise

Deep Learning & Computer Vision

Design and optimization of deep learning models for medical image analysis, especially using CNNs and Transformers.
Expertise in Compact Convolutional Transformers (CCT) and hybrid approaches combining attention mechanisms with convolutional tokenization.
Medical Image Processing and Analysis

Advanced image processing techniques to enhance the quality of retinal images, including median filtering, CLAHE, and noise reduction.
Feature extraction, data augmentation, and class imbalance correction to strengthen model robustness.
Diabetic Retinopathy Classification

Development of AI-based solutions for detecting the stages of diabetic retinopathy, using innovative models like DRCCT and ARNet-DR.
Application of transfer learning with modified architectures such as ResNet50 combined with attention mechanisms to improve classification accuracy.
Security and Digital Watermarking

Integration of AI with digital watermarking to ensure the security and authenticity of medical images, using autoencoder-based architectures.
Design of watermark embedding strategies and enhancement of watermarking systems to ensure robustness against attacks.
Teaching & Training

Experienced university lecturer, training students in artificial intelligence, deep learning, computer vision, and medical image processing.
Proven ability to simplify and teach complex concepts, from neural networks to advanced architectures like Transformers.
Hands-on training in tools and frameworks such as Python, TensorFlow, PyTorch, OCR, CNNs, and GANs, with a focus on real-world problem solving.
Leadership & Senior Expertise

As a Senior AI Consultant and Principal Data Scientist, I have led and participated in projects across various sectors, including finance, healthcare, and information technology.
Project management experience in developing software for automated document processing and anomaly detection in medical signals, especially ECG.
Collaboration with multidisciplinary teams, sharing my expertise to solve complex problems, and ensuring effective communication between developers, researchers, and non-technical stakeholders.
AI & Finance Project Development

Major contributions to the design of AI-based investment platforms, integrating sophisticated algorithms for financial data analysis and automated decision-making.
Innovation & Vision

Visionary in integrating big data, information security, and AI technologies into innovative solutions for the healthcare sector.
Committed to using my skills to improve healthcare and advance scientific research.
Thanks to this multidisciplinary expertise and recognized leadership, I continue to push the boundaries of data science, artificial intelligence, and education, actively contributing to innovations that have a positive impact across various sectors, particularly in healthcare.