Generalization of Pointnet Framework with Synergy of Random Forest Classifier

DOI

10.2991/978-94-6463-529-4_4How to use a DOI?

Keywords

Hybrid models; PointNet; SVM; RF; GBM; KNN

Abstract

In the real world, data is perceived in three dimensions (3D). Automatically analysing 3D visual features is essential in several real-time applications notably autonomous robots, autonomous vehicles, and augmented reality. It is apparent that 3D data are represented in a variety of ways, such as a polygonal mesh, volumetric pixel grid, point cloud, etc., in contrast to 2D images, which are represented as pixel arrays. 90% of current developments in computer vision and machine learning solely pertain to two-dimensional pictures. Point clouds are a popular type of 3D point cloud data (PCD) information captured by visual sensors such as LIDAR and RGB-Depth camera. One of crucial challenges in 3D object classification task while handling the 3D PCD information is the lack of connectivity in 3D space representation. We present a hybrid model in this paper that uses PointNet deep learning model in first phase to automatically extract features as well as reduce the dimension of features and to train a variety of machine learning models in second phase to classify such extracted features. By contrasting several machine learning models according to evaluation metrics, we are able to identify the optimal machine learning model. The presented hybrid model provided higher accuracy than existing method. Moreover, the proposed hybrid model is capable of more accurately classifying the PCD object even with the lesser dense points and computationally efficient in making quicker decisions which suits the essential requirements of real-time autonomous navigation applications.

Copyright

© 2024 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - S. Sridevi
AU  - P. Paayas
AU  - E. Balasubramaniam
AU  - S. Balachandran
AU  - T. Kanimozhi
AU  - M. Shanmugakumar
AU  - Jae Sung Choi
PY  - 2024
DA  - 2024/10/04
TI  - Generalization of Pointnet Framework with Synergy of Random Forest Classifier
BT  - Proceedings of the International Conference on Signal Processing and Computer Vision (SIPCOV-2023)
PB  - Atlantis Press
SP  - 37
EP  - 45
SN  - 2352-5401
UR  - https://doi.org/10.2991/978-94-6463-529-4_4
DO  - 10.2991/978-94-6463-529-4_4
ID  - Sridevi2024
ER  -