Improving Open-Set Semantic Segmentation in 3D Point Clouds by Conditional Channel Capacity Maximization: Preliminary Results

TL;DR

This paper introduces a novel regularizer called Conditional Channel Capacity Maximization (3CM) that enhances open-set semantic segmentation in 3D point clouds by encouraging the model to retain richer, label-dependent features, improving unseen object detection.

Contribution

The paper proposes a plug-and-play framework with a new regularizer, 3CM, based on mutual information, to improve open-set segmentation performance in 3D point clouds.

Findings

3CM improves detection of unseen objects in 3D point clouds.

Incorporating 3CM into standard loss functions enhances feature richness and class distinction.

Experimental results show significant gains in open-set segmentation accuracy.

Abstract

Point-cloud semantic segmentation underpins a wide range of critical applications. Although recent deep architectures and large-scale datasets have driven impressive closed-set performance, these models struggle to recognize or properly segment objects outside their training classes. This gap has sparked interest in Open-Set Semantic Segmentation (O3S), where models must both correctly label known categories and detect novel, unseen classes. In this paper, we propose a plug and play framework for O3S. By modeling the segmentation pipeline as a conditional Markov chain, we derive a novel regularizer term dubbed Conditional Channel Capacity Maximization (3CM), that maximizes the mutual information between features and predictions conditioned on each class. When incorporated into standard loss functions, 3CM encourages the encoder to retain richer, label-dependent features, thereby enhancing the network's ability to distinguish and segment previously unseen categories. Experimental results demonstrate effectiveness of proposed method on detecting unseen objects. We further outline future directions for dynamic open-world adaptation and efficient information-theoretic estimation.