Clément Rambour

RDFIA (M2 SU)

RDFIA – Pattern Recognition for Image Analysis and Interpretation

Objective: This course introduces key concepts and methods for the automatic analysis and interpretation of visual content in images. Modern machine learning approaches are used to explore fundamental and advanced methods for computer vision, with emphasis on deep learning architectures and their training.

The course covers image classification, segmentation, vision–language models, and generative modeling. Topics include convolutional neural networks, Vision Transformers, self-supervised learning, diffusion models, robustness, explainability, transfer learning, and domain adaptation.

Theoretical lectures are complemented by hands-on Python practicals where students implement and experiment with the studied models. Practicals and code resources: Here.

Course materials and indicative timeline (subject to change)

  • Course 1 (Sep. 23, 2026) — Introduction to Computer Vision and ML basics: slides
  • Course 2 (Sep. 30, 2026) — Introduction to Neural Networks and training: slides
  • Course 3 (Oct. 07, 2026) — Datasets, benchmarks and evaluation; Neural Nets for Image Classification: data slides, classification slides
  • Course 4 (Oct. 14, 2026) — Convolutional architectures and normalization: slides
  • Course 5 (Oct. 21, 2026) — Large ConvNets and Vision Transformers: large convnets, ViT
  • Course 6 (Nov. 04, 2026) — Vision Transformers details; Transfer learning and Domain adaptation: slides
  • Course 7 (Nov. 18, 2026) — Vision–Language Models (CLIP, MLLMs) and explainability: CLIP, MLLM
  • Course 8 (Dec. 02, 2026) — From Classification to Segmentation using ConvNets and Transformers: segmentation, SAM/transformer segmentation
  • Course 9 (Dec. 09, 2026) — Self-Supervised Learning in Vision and Multimodality: slides, and SAM (family)
  • Course 10 (Dec. 16, 2026) — Generative models with GANs and conditional GANs/diffusion: GAN, cGAN / diffusion
  • Course 11 (Jan. 06, 2027) — Control and practical sessions (dates and times announced in class)
  • Course 12 (Jan. 13, 2027) — Diffusion models for image generation: slides
  • Course 13 (Jan. 20, 2027) — Bayesian deep learning
  • Course 14 (Jan. 27, 2027) — Failure and out-of-distribution detection

Prerequisites

Basic knowledge of digital image representation, statistical data processing, and scientific computing in Python.

Further reading