Complementary Bi-directional Feature Compression for Indoor 360{\deg} Semantic Segmentation with Self-distillation

TL;DR

This paper introduces a novel 360-degree semantic segmentation method that combines bi-directional feature compression with self-distillation, effectively leveraging complementary representations to improve accuracy and visual quality.

Contribution

The proposed approach uniquely combines vertical and horizontal feature compression with a self-distillation strategy for enhanced panoramic semantic segmentation.

Findings

Achieves at least 10% improvement over state-of-the-art methods.

Outperforms existing solutions in quantitative evaluations.

Displays superior visual appearance in segmentation results.

Abstract

Recently, horizontal representation-based panoramic semantic segmentation approaches outperform projection-based solutions, because the distortions can be effectively removed by compressing the spherical data in the vertical direction. However, these methods ignore the distortion distribution prior and are limited to unbalanced receptive fields, e.g., the receptive fields are sufficient in the vertical direction and insufficient in the horizontal direction. Differently, a vertical representation compressed in another direction can offer implicit distortion prior and enlarge horizontal receptive fields. In this paper, we combine the two different representations and propose a novel 360{\deg} semantic segmentation solution from a complementary perspective. Our network comprises three modules: a feature extraction module, a bi-directional compression module, and an ensemble decoding module. First, we extract multi-scale features from a panorama. Then, a bi-directional compression module is designed to compress features into two complementary low-dimensional representations, which provide content perception and distortion prior. Furthermore, to facilitate the fusion of bi-directional features, we design a unique self distillation strategy in the ensemble decoding module to enhance the interaction of different features and further improve the performance. Experimental results show that our approach outperforms the state-of-the-art solutions with at least 10\% improvement on quantitative evaluations while displaying the best performance on visual appearance.