Real-time scene reconstruction is an important task in various real-life applications such as self-driving, AR/VR, robotic navigation, and intelligent manufacture. This project aims to provide opportunities for the applicants to research and implement a deep-learning-based scene reconstruction framework, which will not only help the applicants to gain experience and knowledge on the topic but also lead to the development of their research skills in the field of computer vision.
Applicants will study the state-of-the-art model for 3D scene reconstruction. They will read the NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis paper and implement it by themselves by referring to and studying the paper. Also, they will analyze the novel contribution of the paper to understand how it outperforms the other existing models.
- Reading the literature on this research problem to get an overall understanding of the task
- Get to know implementation details in existing representative works
- Implement a deep learning-based framework for the task