3D Computer Vision is a vibrant field of research, centered around creating advanced computational methodologies to interpret the complex reality of our surrounding environment, inherently three-dimensional and continually changing over time. Over the recent years, the advent of learning methodologies, particularly those based on neural networks, have served as a catalyst for swift advancement in this discipline, advancing the state-of-the-art across numerous 3D tasks, including reconstruction, semantic scene understanding, object detection, and content generation. In this thought-provoking seminar, we will embark on a journey exploring seminal works that cover a wide spectrum of learning methods specifically tailored for 3D applications. Topics to be covered include: supervised and unsupervised learning for 3D shapes and scenes, test-time optimization, neural rendering, generative models, dynamic representations, among others. The content will provide attendees with a comprehensive understanding of current techniques and inspire them to contribute to further advancements in the field of 3D Computer Vision
- מורה: אור ליטני