17714347. METHOD AND APPARATUS WITH DEPTH MAP GENERATION simplified abstract (Samsung Electronics Co., Ltd.)
Contents
METHOD AND APPARATUS WITH DEPTH MAP GENERATION
Organization Name
Inventor(s)
Seokhwan Jang of Anyang-si (KR)
METHOD AND APPARATUS WITH DEPTH MAP GENERATION - A simplified explanation of the abstract
This abstract first appeared for US patent application 17714347 titled 'METHOD AND APPARATUS WITH DEPTH MAP GENERATION
Simplified Explanation
The patent application describes a method and apparatus for generating depth maps using multi-view depth maps and camera parameters. Here is a simplified explanation of the abstract:
- The method involves generating points for a point cloud by converting multi-view depth maps into a 3D space using camera parameters.
- Feature embedding vectors are extracted for the generated points.
- A 2D feature map is generated for a specific view based on the extracted feature embedding vectors.
- A refinement network is used to generate a residual depth map based on the 2D feature map.
- A new depth map is generated by combining the residual depth map with an initial depth map from the multi-view depth maps.
Potential Applications
This technology has potential applications in various fields, including:
- Computer vision: The generated depth maps can be used for object recognition, scene understanding, and augmented reality applications.
- Robotics: Accurate depth maps can aid in robot navigation, obstacle avoidance, and object manipulation tasks.
- Autonomous vehicles: Depth maps can enhance the perception capabilities of autonomous vehicles, enabling better object detection and tracking.
Problems Solved
The method and apparatus described in the patent application address the following problems:
- Depth map generation: The technology provides a method to generate accurate depth maps from multi-view depth maps, improving the understanding of the 3D structure of a scene.
- Feature extraction: The method extracts feature embedding vectors to represent the points in the generated point cloud, enabling efficient analysis and processing of the data.
- Depth map refinement: The use of a refinement network helps to enhance the quality of the depth maps, reducing noise and improving accuracy.
Benefits
The use of this technology offers several benefits:
- Improved depth map accuracy: By combining multiple views and utilizing camera parameters, the generated depth maps are more accurate and reliable.
- Efficient processing: The extraction of feature embedding vectors and the use of a refinement network optimize the processing of depth map generation.
- Versatility: The method can be applied to various scenarios and applications, providing flexibility in utilizing depth maps for different purposes.
Original Abstract Submitted
A method and apparatus with depth map generation. The method may include generating points for a point cloud by unprojecting multi-view depth maps, of plural views, into a corresponding three-dimensional (3D) space using respective camera parameters corresponding to each view of the multi-view depth maps, extracting feature embedding vectors corresponding to the generated points, generating a two-dimensional (2D) feature map of a set view based on the extracted feature embedding vectors, generating a residual depth map using a refinement network with respect to the 2D feature map, generating a new depth map based on the residual depth map and an initial depth map, of the set view, among the multi-view depth maps.
