RDFIA: Reconnaissance Des Formes par Intelligence Artificielle (Recognition and Description of Patterns with Artificial Intelligence)

This repository contains practical works (PW) I completed (with Nathan Galmiche) as part of my Master's course in deep learning for visual understanding (https://cord.isir.upmc.fr/teaching-rdfia/).

You can take a look at the reports we wrote by clicking to this link: https://github.com/pictoune/RDFIA/blob/main/all_reports.pdf.

Table of Contents

• PW 1: SIFT & BOW, SVM
• PW 2: Transfer Learning & Visualization, Domain Adaptation & GAN
• PW 3: Advanced Topics in Transfer Learning and GANs
• PW 4: Bayesian Methods and Uncertainty in Deep Learning

PW 1: SIFT & BOW, SVM

This section contains Python code (Jupyter notebook) for image processing with SIFT & BOW and SVM techniques, and a report presenting the results.

• Code: SIFT & BOW, SVM
• Report: PDF

PW 2: Transfer Learning & Visualization, Domain Adaptation & GAN

This section contains Python code (Jupyter notebook) for exploring transfer learning, visualization, domain adaptation, and GANs, and a report presenting the results.

• Code: Transfer Learning & Visualization, Domain Adaptation & GAN
• Report: PDF

PW 3: Advanced Topics in Transfer Learning and GANs

This section contains Python code (Jupyter notebook) for advanced topics in transfer learning and GANs, and a report presenting the results.

• Code: Transfer Learning, Visualizing Neural Networks, Domain Adaptation, GANs
• Report: PDF

PW 4: Bayesian Methods and Uncertainty in Deep Learning

This section contains Python code (Jupyter notebook) for Bayesian methods and uncertainty in deep learning, and a report presenting the results.

• Code: Bayesian Linear Regression, Approximate Inference, Uncertainty Applications
• Report: PDF

License

This project is open source and available under the MIT License.

Feel free to explore the projects and reach out if you have any questions or suggestions.