3D Point Cloud Classification and Segmentation using 3D Modified Fisher Vector Representation for Convolutional Neural Networks

TL;DR

This paper introduces a novel 3D point cloud representation called 3D Modified Fisher Vectors (3DmFV) that combines grid structure with Fisher vectors, enabling efficient CNN-based classification and segmentation with competitive results.

Contribution

The paper proposes a new hybrid 3D point cloud representation, 3DmFV, and a CNN architecture tailored for it, improving efficiency and performance over existing methods.

Findings

Achieves competitive or superior performance on benchmark datasets.

Provides a compact and computationally efficient representation.

Demonstrates effectiveness in classification and segmentation tasks.

Abstract

The point cloud is gaining prominence as a method for representing 3D shapes, but its irregular format poses a challenge for deep learning methods. The common solution of transforming the data into a 3D voxel grid introduces its own challenges, mainly large memory size. In this paper we propose a novel 3D point cloud representation called 3D Modified Fisher Vectors (3DmFV). Our representation is hybrid as it combines the discrete structure of a grid with continuous generalization of Fisher vectors, in a compact and computationally efficient way. Using the grid enables us to design a new CNN architecture for point cloud classification and part segmentation. In a series of experiments we demonstrate competitive performance or even better than state-of-the-art on challenging benchmark datasets.