Comprehensive Study and Analysis of Point Transformer for Point Cloud Data

Abstract

Self-attention networks are making significant progress in picture analysis tasks like image categorization and object detection, just as they have revolutionized natural language processing. This paper represents the use of self-attention networks for processing point cloud data in response to this achievement. To build self-attention networks for tasks like object categorization, this paper has explored self-attention layers for point clouds. It is used in many applications such as Point Cloud Classification, Object Detection, Point Cloud Generation etc. This research paper covers the basics of Point Transformer and its importance in Point Cloud Data. the performance of the Point Transformer algorithm was evaluated on benchmark datasets, such as ModelNet10 with 92.40% accuracy.