PointMixer: MLP-Mixer for Point Cloud Understanding

22 Nov 2021  ·  Jaesung Choe, Chunghyun Park, Francois Rameau, Jaesik Park, In So Kweon ·

MLP-Mixer has newly appeared as a new challenger against the realm of CNNs and transformer. Despite its simplicity compared to transformer, the concept of channel-mixing MLPs and token-mixing MLPs achieves noticeable performance in visual recognition tasks. Unlike images, point clouds are inherently sparse, unordered and irregular, which limits the direct use of MLP-Mixer for point cloud understanding. In this paper, we propose PointMixer, a universal point set operator that facilitates information sharing among unstructured 3D points. By simply replacing token-mixing MLPs with a softmax function, PointMixer can "mix" features within/between point sets. By doing so, PointMixer can be broadly used in the network as inter-set mixing, intra-set mixing, and pyramid mixing. Extensive experiments show the competitive or superior performance of PointMixer in semantic segmentation, classification, and point reconstruction against transformer-based methods.

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
3D Point Cloud Classification ModelNet40 PointMixer Overall Accuracy 93.6 # 41
Mean Accuracy 91.4 # 11
Number of params 6.5M # 98
Semantic Segmentation S3DIS Area5 PointMixer mIoU 71.4 # 18
mAcc 77.4 # 16
Number of params 6.5M # 45