PSFormer: Point Transformer for 3D Salient Object Detection

TL;DR

PSFormer is a novel point transformer model designed for 3D salient object detection, effectively capturing multi-scale contextual information at point and scene levels to improve detection accuracy.

Contribution

The paper introduces PSFormer, a new encoder-decoder architecture with specialized transformer modules for point and scene context modeling in 3D object detection.

Findings

Outperforms existing methods in 3D salient object detection

More robust to small, multiple, and complex objects

Effective multi-scale context modeling

Abstract

We propose PSFormer, an effective point transformer model for 3D salient object detection. PSFormer is an encoder-decoder network that takes full advantage of transformers to model the contextual information in both multi-scale point- and scene-wise manners. In the encoder, we develop a Point Context Transformer (PCT) module to capture region contextual features at the point level; PCT contains two different transformers to excavate the relationship among points. In the decoder, we develop a Scene Context Transformer (SCT) module to learn context representations at the scene level; SCT contains both Upsampling-and-Transformer blocks and Multi-context Aggregation units to integrate the global semantic and multi-level features from the encoder into the global scene context. Experiments show clear improvements of PSFormer over its competitors and validate that PSFormer is more robust to challenging cases such as small objects, multiple objects, and objects with complex structures.