View source for 20230094308. DATASET GENERATION METHOD FOR SELF-SUPERVISED LEARNING SCENE POINT CLOUD COMPLETION BASED ON PANORAMAS simplified abstract (DALIAN UNIVERSITY OF TECHNOLOGY)

=DATASET GENERATION METHOD FOR SELF-SUPERVISED LEARNING SCENE POINT CLOUD COMPLETION BASED ON PANORAMAS=


==Organization Name==

[[:Category:DALIAN UNIVERSITY OF TECHNOLOGY|DALIAN UNIVERSITY OF TECHNOLOGY]]


[[Category:DALIAN UNIVERSITY OF TECHNOLOGY]]

==Inventor(s)==

[[:Category:Xin Yang of Dalian (CN)|Xin Yang of Dalian (CN)]][[Category:Xin Yang of Dalian (CN)]]

[[:Category:Tong Li of Dalian (CN)|Tong Li of Dalian (CN)]][[Category:Tong Li of Dalian (CN)]]

[[:Category:Baocai Yin of Dalian (CN)|Baocai Yin of Dalian (CN)]][[Category:Baocai Yin of Dalian (CN)]]

[[:Category:Zhaoxuan Zhang of Dalian (CN)|Zhaoxuan Zhang of Dalian (CN)]][[Category:Zhaoxuan Zhang of Dalian (CN)]]

[[:Category:Boyan Wei of Dalian (CN)|Boyan Wei of Dalian (CN)]][[Category:Boyan Wei of Dalian (CN)]]

[[:Category:Zhenjun Du of Dalian (CN)|Zhenjun Du of Dalian (CN)]][[Category:Zhenjun Du of Dalian (CN)]]

==DATASET GENERATION METHOD FOR SELF-SUPERVISED LEARNING SCENE POINT CLOUD COMPLETION BASED ON PANORAMAS - A simplified explanation of the abstract==


This abstract first appeared for US patent application 20230094308 titled 'DATASET GENERATION METHOD FOR SELF-SUPERVISED LEARNING SCENE POINT CLOUD COMPLETION BASED ON PANORAMAS

==Simplified Explanation==

The present invention is a dataset generation method for self-supervised learning scene point cloud completion based on panoramas in the field of computer vision. It involves generating pairs of incomplete point cloud and target point cloud with RGB and normal information by using RGB panoramas, depth panoramas, and normal panoramas as input. This dataset is used to train the scene point cloud completion network.

* Simplification of the collection mode of point cloud data in a real scene
* Occlusion prediction and equirectangular projection based on view conversion
* Processing of the stripe problem and point-to-point occlusion problem during conversion
* Design of view selection strategy

==Potential Applications==
* Computer vision applications that require 3D reconstruction
* Augmented reality and virtual reality applications
* Robotics and autonomous navigation systems
* Environmental monitoring and mapping

==Problems Solved==
* Simplifies the collection of point cloud data in real scenes, making it more efficient and cost-effective
* Addresses the challenges of occlusion prediction and view conversion in generating complete point clouds
* Solves the stripe problem and point-to-point occlusion problem during the conversion process

==Benefits==
* Enables the generation of a self-supervised learning dataset for training scene point cloud completion networks
* Improves the accuracy and completeness of reconstructed 3D scenes
* Reduces the time and effort required for 3D reconstruction in computer vision applications
* Enhances the performance of computer vision systems in various domains


==Original Abstract Submitted==

the present invention belongs to the technical field of 3d reconstruction in the field of computer vision, and provides a dataset generation method for self-supervised learning scene point cloud completion based on panoramas. pairs of incomplete point cloud and target point cloud with rgb information and normal information can be generated by taking rgb panoramas, depth panoramas and normal panoramas in the same view as input for constructing a self-supervised learning dataset for training of the scene point cloud completion network. the key points of the present invention are occlusion prediction and equirectangular projection based on view conversion, and processing of the stripe problem and point-to-point occlusion problem during conversion. the method of the present invention includes simplification of the collection mode of the point cloud data in a real scene; occlusion prediction idea of view conversion; and design of view selection strategy.



[[Category:G06V20/64]]
[[Category:G06T7/50]]
[[Category:G06T3/00]]