RWKV-PCSSC: Exploring RWKV Model for Point Cloud Semantic Scene Completion

2025/11/13 by He, Wenzhe, Chen, Xiaojun, Chen, Wentang +3
#Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences

paper · doi:10.48550/arxiv.2511.09878

Abstract

Semantic Scene Completion (SSC) aims to generate a complete semantic scene from an incomplete input. Existing approaches often employ dense network architectures with a high parameter count, leading to increased model complexity and resource demands. To address these limitations, we propose RWKV-PCSSC, a lightweight point cloud semantic scene completion network inspired by the Receptance Weighted Key Value (RWKV) mechanism. Specifically, we introduce a RWKV Seed Generator (RWKV-SG) module that can aggregate features from a partial point cloud to produce a coarse point cloud with coarse features. Subsequently, the point-wise feature of the point cloud is progressively restored through multiple stages of the RWKV Point Deconvolution (RWKV-PD) modules. By leveraging a compact and efficient design, our method achieves a lightweight model representation. Experimental results demonstrate that RWKV-PCSSC reduces the parameter count by 4.18× and improves memory efficiency by 1.37× compared to state-of-the-art methods PointSSC. Furthermore, our network achieves state-of-the-art performance on established indoor (SSC-PC, NYUCAD-PC) and outdoor (PointSSC) scene dataset, as well as on our proposed datasets (NYUCAD-PC-V2, 3D-FRONT-PC).

Citations

• OccRWKV: Rethinking Efficient 3D Semantic Occupancy Prediction with Linear Complexity
• PointRWKV: Efficient RWKV-Like Model for Hierarchical Point Cloud Learning
• KPConvX: Modernizing Kernel Point Convolution with Kernel Attention
• Mamba3D: Enhancing Local Features for 3D Point Cloud Analysis via State Space Model
• Diffusion-RWKV: Scaling RWKV-Like Architectures for Diffusion Models
• Vision-RWKV: Efficient and Scalable Visual Perception with RWKV-Like Architectures
• PointMamba: A Simple State Space Model for Point Cloud Analysis
• Point Transformer V3: Simpler, Faster, Stronger
• RWKV: Reinventing RNNs for the Transformer Era
• Semantic Scene Completion with Cleaner Self
• AdaPoinTr: Diverse Point Cloud Completion with Adaptive Geometry-Aware Transformers
• AdaPoinTr: Diverse Point Cloud Completion With Adaptive Geometry-Aware Transformers
• CasFusionNet: A Cascaded Network for Point Cloud Semantic Scene Completion by Dense Feature Fusion
• Point Transformer V2: Grouped Vector Attention and Partition-based Pooling
• SeedFormer: Patch Seeds based Point Cloud Completion with Upsample Transformer
• PMP-Net++: Point Cloud Completion by Transformer-Enhanced Multi-step Point Moving Paths
• Snowflake Point Deconvolution for Point Cloud Completion and Generation with Skip-Transformer
• Rethinking Network Design and Local Geometry in Point Cloud: A Simple Residual MLP Framework
• Data Augmented 3D Semantic Scene Completion with 2D Segmentation Priors
• SSCNav: Confidence-Aware Semantic Scene Completion for Visual Semantic Navigation
• PMP-Net: Point Cloud Completion by Learning Multi-step Point Moving Paths
• 3D-FRONT: 3D Furnished Rooms with layOuts and semaNTics
• 3D-FUTURE: 3D Furniture shape with TextURE
• Anisotropic Convolutional Networks for 3D Semantic Scene Completion
• SqueezeSegV3: Spatially-Adaptive Convolution for Efficient Point-Cloud Segmentation
• 3D Sketch-aware Semantic Scene Completion via Semi-supervised Structure Prior
• Morphing and Sampling Network for Dense Point Cloud Completion
• KPConv: Flexible and Deformable Convolution for Point Clouds
• Efficient Semantic Scene Completion Network with Spatial Group Convolution
• RGCNN: Regularized Graph CNN for Point Cloud Segmentation
• Dynamic Graph CNN for Learning on Point Clouds
• Dynamic Graph CNN for Learning on Point Clouds
• PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space
• Cognitive Mapping and Planning for Visual Navigation
• PointNet: Deep Learning on Point Sets for 3D Classification and\n Segmentation
• Semantic Scene Completion from a Single Depth Image
• Predicting Complete 3D Models of Indoor Scenes

Related

• CasFusionNet: A Cascaded Network for Point Cloud Semantic Scene Completion by Dense Feature Fusion
• Learning Local Displacements for Point Cloud Completion
• SCPNet: Semantic Scene Completion on Point Cloud
• PointSSC: A Cooperative Vehicle-Infrastructure Point Cloud Benchmark for Semantic Scene Completion
• Monocular Semantic Scene Completion via Masked Recurrent Networks