# Flex-Convolution (Million-Scale Point-Cloud Learning Beyond Grid-Worlds)

## TL;DR

This paper introduces Flex-Convolution, a generalized convolution layer designed for unstructured 3D point clouds, enabling efficient processing of millions of points with improved performance and lower resource consumption.

## Contribution

It presents a novel flex-convolution layer and GPU implementation that effectively handles large-scale point clouds beyond grid-based data assumptions.

## Key findings

- Competitive results on small benchmarks with fewer parameters.
- Significant improvements on million-scale real-world datasets.
- First method to efficiently process 7 million points concurrently.

## Abstract

Traditional convolution layers are specifically designed to exploit the natural data representation of images -- a fixed and regular grid. However, unstructured data like 3D point clouds containing irregular neighborhoods constantly breaks the grid-based data assumption. Therefore applying best-practices and design choices from 2D-image learning methods towards processing point clouds are not readily possible. In this work, we introduce a natural generalization flex-convolution of the conventional convolution layer along with an efficient GPU implementation. We demonstrate competitive performance on rather small benchmark sets using fewer parameters and lower memory consumption and obtain significant improvements on a million-scale real-world dataset. Ours is the first which allows to efficiently process 7 million points concurrently.