DANet: Density Adaptive Convolutional Network with Interactive Attention for 3D Point Clouds

TL;DR

This paper introduces DAConv and IAM, novel modules for 3D point cloud analysis that adapt to varying densities and efficiently model contextual dependencies, achieving state-of-the-art results.

Contribution

The paper presents density adaptive convolution (DAConv) and an interactive attention module (IAM) that improve robustness and efficiency in 3D point cloud processing.

Findings

DAConv is more robust to point density variations.

Achieves 93.6% accuracy on ModelNet40.

Competitive results on S3DIS and ShapeNet datasets.

Abstract

Local features and contextual dependencies are crucial for 3D point cloud analysis. Many works have been devoted to designing better local convolutional kernels that exploit the contextual dependencies. However, current point convolutions lack robustness to varying point cloud density. Moreover, contextual modeling is dominated by non-local or self-attention models which are computationally expensive. To solve these problems, we propose density adaptive convolution, coined DAConv. The key idea is to adaptively learn the convolutional weights from geometric connections obtained from the point density and position. To extract precise context dependencies with fewer computations, we propose an interactive attention module (IAM) that embeds spatial information into channel attention along different spatial directions. DAConv and IAM are integrated in a hierarchical network architecture to achieve local density and contextual direction-aware learning for point cloud analysis. Experiments show that DAConv is significantly more robust to point density compared to existing methods and extensive comparisons on challenging 3D point cloud datasets show that our network achieves state-of-the-art classification results of 93.6% on ModelNet40, competitive semantic segmentation results of 68.71% mIoU on S3DIS and part segmentation results of 86.7% mIoU on ShapeNet.