Explore image domain transfer techniques in this NVIDIA guest lecture from MIT's Introduction to Deep Learning course. Delve into artistic style transfer, photo smoothing, and various neural network approaches for transforming images across different domains. Learn about unimodal and multimodal translations, temporal consistency, and applications such as converting winter scenes to summer. Gain insights into the latest advancements in deep learning for image manipulation and understand the potential of these techniques through user studies and practical examples.
MIT: Image Domain Transfer - NVIDIA Guest Lecture
Alexander Amini and Massachusetts Institute of Technology via YouTube
Overview
Syllabus
Introduction
Image Domain Transfer
Artistic Style Transfer
Labels
Photo Smoothing
Comparisons
User Study
Neural Networks
Unimodal multimodal
Fixed to HD
Conditional
Multimodal
Temporal Consistency
Edges
Poses
Game Engine
Neural Network
Shared latent space
Winter to Summer
Unimodal Translation
Multimodal Translation
Conclusion
Thank you
Taught by
https://www.youtube.com/@AAmini/videos
