CTFS : Collaborative Teacher Framework for Forward-Looking Sonar Image Semantic Segmentation with Extremely Limited Labels

2026-03-22

AI Summary

This work addresses the challenges of semantic segmentation in forward-looking sonar (FLS) imagery under extremely low annotation rates, where factors such as speckle noise, low texture contrast, acoustic shadows, and geometric distortions hinder effective learning within conventional teacher–student frameworks. To overcome these limitations, the authors propose a collaborative teacher framework that integrates a general-purpose teacher with multiple sonar-specific teachers. This architecture employs alternating multi-teacher guidance and a cross-teacher consistency evaluation mechanism to dynamically assess the reliability of pseudo-labels and suppress noise-induced errors. The approach substantially enhances the student model’s capacity to capture sonar-specific features, achieving a 5.08% absolute improvement in mean Intersection over Union (mIoU) over the current state-of-the-art method on the FLSMD dataset with only 2% labeled data, thereby enabling more robust semi-supervised semantic segmentation.

Technology Category

• Machine Learning: Semi-Supervised Learning
• Computer Vision: Segmentation
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

As one of the most important underwater sensing technologies, forward-looking sonar exhibits unique imaging characteristics. Sonar images are often affected by severe speckle noise, low texture contrast, acoustic shadows, and geometric distortions. These factors make it difficult for traditional teacher-student frameworks to achieve satisfactory performance in sonar semantic segmentation tasks under extremely limited labeled data conditions. To address this issue, we propose a Collaborative Teacher Semantic Segmentation Framework for forward-looking sonar images. This framework introduces a multi-teacher collaborative mechanism composed of one general teacher and multiple sonar-specific teachers. By adopting a multi-teacher alternating guidance strategy, the student model can learn general semantic representations while simultaneously capturing the unique characteristics of sonar images, thereby achieving more comprehensive and robust feature modeling. Considering the challenges of sonar images, which can lead teachers to generate a large number of noisy pseudo-labels, we further design a cross-teacher reliability assessment mechanism. This mechanism dynamically quantifies the reliability of pseudo-labels by evaluating the consistency and stability of predictions across multiple views and multiple teachers, thereby mitigating the negative impact caused by noisy pseudo-labels. Notably, on the FLSMD dataset, when only 2% of the data is labeled, our method achieves a 5.08% improvement in mIoU compared to other state-of-the-art approaches.

Problem

• forward-looking sonar
• semantic segmentation
• limited labels
• speckle noise
• pseudo-labels

Innovation

• Collaborative Teacher Framework
• Forward-Looking Sonar
• Semantic Segmentation
• Limited Labels
• Pseudo-Label Reliability