# Escape from Cells: Deep Kd-Networks for the Recognition of 3D Point Cloud Models

## TL;DR

This paper introduces Kd-networks, a novel deep learning architecture for 3D point cloud recognition that avoids grid-based limitations and demonstrates competitive performance on shape classification, retrieval, and segmentation tasks.

## Contribution

The paper proposes Kd-networks, a new deep learning model that processes unstructured 3D point clouds efficiently without relying on grid rasterization, improving scalability and performance.

## Key findings

- Kd-networks perform competitively on shape recognition benchmarks.
- They effectively handle unstructured point cloud data.
- The architecture avoids the poor scaling issues of grid-based methods.

## Abstract

We present a new deep learning architecture (called Kd-network) that is designed for 3D model recognition tasks and works with unstructured point clouds. The new architecture performs multiplicative transformations and share parameters of these transformations according to the subdivisions of the point clouds imposed onto them by Kd-trees. Unlike the currently dominant convolutional architectures that usually require rasterization on uniform two-dimensional or three-dimensional grids, Kd-networks do not rely on such grids in any way and therefore avoid poor scaling behaviour. In a series of experiments with popular shape recognition benchmarks, Kd-networks demonstrate competitive performance in a number of shape recognition tasks such as shape classification, shape retrieval and shape part segmentation.