Snap inc. (20240273809). 3D MODELING BASED ON NEURAL LIGHT FIELD simplified abstract
Contents
3D MODELING BASED ON NEURAL LIGHT FIELD
Organization Name
Inventor(s)
Zeng Huang of Los Angeles CA (US)
Jian Ren of Marina Del Ray CA (US)
Sergey Tulyakov of Santa Monica CA (US)
Menglei Chai of Los Angeles CA (US)
Kyle Olszewski of Los Angeles CA (US)
Huan Wang of Somerville MA (US)
3D MODELING BASED ON NEURAL LIGHT FIELD - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240273809 titled '3D MODELING BASED ON NEURAL LIGHT FIELD
The patent application describes methods and systems for creating a 3D model of a scene using machine learning technology.
- Receiving a set of 2D images of a real-world environment.
- Applying a neural light field network to predict pixel values of a target image representing a different view of the environment.
- Training the machine learning model to map ray origin and direction to pixel values.
- Generating a 3D model based on the 2D images and predicted target image.
Potential Applications: - Virtual reality and augmented reality development. - Architectural design and visualization. - Gaming and simulation industries.
Problems Solved: - Enhancing the accuracy and efficiency of 3D modeling. - Streamlining the process of creating realistic virtual environments.
Benefits: - Improved realism in virtual environments. - Faster creation of 3D models. - Enhanced user experience in VR and AR applications.
Commercial Applications: Title: Advanced 3D Modeling Technology for Virtual Environments This technology can be used in industries such as gaming, architecture, and virtual reality development to create realistic 3D models efficiently, leading to enhanced user experiences and improved design processes.
Questions about 3D Modeling Technology: 1. How does this technology improve the accuracy of 3D models compared to traditional methods? 2. What are the potential cost-saving benefits for industries utilizing this advanced 3D modeling technology?
Original Abstract Submitted
methods and systems are disclosed for performing operations for generating a 3d model of a scene. the operations include: receiving a set of two-dimensional (2d) images representing a first view of a real-world environment; applying a machine learning model comprising a neural light field network to the set of 2d images to predict pixel values of a target image representing a second view of the real-world environment, the machine learning model being trained to map a ray origin and direction directly to a given pixel value; and generating a three-dimensional (3d) model of the real-world environment based on the set of 2d images and the predicted target image.
