RL-AD-Net: Reinforcement Learning Guided Adaptive Displacement in Latent Space for Refined Point Cloud Completion

TL;DR

RL-AD-Net enhances point cloud completion by using reinforcement learning to refine latent space representations, improving local geometric details while maintaining global shape plausibility across diverse categories.

Contribution

It introduces a reinforcement learning framework operating in latent space for adaptive refinement of point cloud completions, a novel approach in this domain.

Findings

Consistently improves completion quality in diverse cropping scenarios.

Operates in a lightweight, model-agnostic manner without retraining existing networks.

Enhances local geometric fidelity while preserving global shape plausibility.

Abstract

Recent point cloud completion models, including transformer-based, denoising-based, and other state-of-the-art approaches, generate globally plausible shapes from partial inputs but often leave local geometric inconsistencies. We propose RL-AD-Net, a reinforcement learning (RL) refinement framework that operates in the latent space of a pretrained point autoencoder. The autoencoder encodes completions into compact global feature vectors (GFVs), which are selectively adjusted by an RL agent to improve geometric fidelity. To ensure robustness, a lightweight non-parametric PointNN selector evaluates the geometric consistency of both the original completion and the RL-refined output, retaining the better reconstruction. When ground truth is available, both Chamfer Distance and geometric consistency metrics guide refinement. Training is performed separately per category, since the unsupervised and dynamic nature of RL makes convergence across highly diverse categories challenging. Nevertheless, the framework can be extended to multi-category refinement in future work. Experiments on ShapeNetCore-2048 demonstrate that while baseline completion networks perform reasonable under their training-style cropping, they struggle in random cropping scenarios. In contrast, RL-AD-Net consistently delivers improvements across both settings, highlighting the effectiveness of RL-guided ensemble refinement. The approach is lightweight, modular, and model-agnostic, making it applicable to a wide range of completion networks without requiring retraining.