Vision, image and machine learning (partie AM)

The foundations of Deep Learning for images #

• Alexandre Meyer

Courses #

• CM1: Introduction to neural networks (forward and backpropagation) (only ID3D)
• CM2: Convolution Neural Networks (CNN)
• CM3: introduction to generative approaches like AE and GAN
• CM4: the “modern” vision (Segmentation, Tracking-YOLO, Transformer for Images, Skeleton Capture)

Lab (TP) #

1. [Only ID3D] A neural network from scratch. (Python, numpy)
2. Style transfer between images (pytorch). Slides.
3. Image classification and attention (pytorch)
4. Gesture transfer and person image generation. (Pytorch, mediapipe)
5. (Only ID3D) TP GAN simple. (Pytorch)

Extra #

0. Installation
1. Classification
   1. 2D Point Classification with Pytorch.
   2. Image classification using Convolution Neural Networks with Pytorch. (CNN, Pytorch)

• Old TP: autoencoder and animation