Investigating Location-Regularised Self-Supervised Feature Learning for Seafloor Visual Imagery

2025-09-08

AI Summary

This study addresses the challenge of enhancing self-supervised learning (SSL) for seafloor image analysis by leveraging geographic metadata. We propose a location-aware regularization method that systematically integrates spatial priors into SSL objectives—applied across multiple frameworks (e.g., SimCLR, BYOL) and architectures (CNNs and Vision Transformers). Crucially, our approach explicitly encodes geographic constraints into the contrastive or predictive loss functions, thereby guiding representation learning to capture spatial correlations inherent in benthic imagery. Evaluated on three diverse, multi-source seafloor image datasets, the method demonstrates consistent improvements: average F1-scores increase by 4.9% for CNNs and 6.3% for ViTs on downstream classification tasks. Notably, lightweight ViT variants trained with geographic regularization achieve performance comparable to fully supervised baselines. These results validate that incorporating geospatial inductive bias significantly improves feature discriminability and generalization in marine visual understanding—particularly under data-limited and label-scarce conditions.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Machine Learning: Unsupervised & Self-Supervised Learning
• Search and Optimization: Learning to Search

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
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

High-throughput interpretation of robotically gathered seafloor visual imagery can increase the efficiency of marine monitoring and exploration. Although recent research has suggested that location metadata can enhance self-supervised feature learning (SSL), its benefits across different SSL strategies, models and seafloor image datasets are underexplored. This study evaluates the impact of location-based regularisation on six state-of-the-art SSL frameworks, which include Convolutional Neural Network (CNN) and Vision Transformer (ViT) models with varying latent-space dimensionality. Evaluation across three diverse seafloor image datasets finds that location-regularisation consistently improves downstream classification performance over standard SSL, with average F1-score gains of $4.9 pm 4.0%$ for CNNs and $6.3 pm 8.9%$ for ViTs, respectively. While CNNs pretrained on generic datasets benefit from high-dimensional latent representations, dataset-optimised SSL achieves similar performance across the high (512) and low (128) dimensional latent representations. Location-regularised SSL improves CNN performance over pre-trained models by $2.7 pm 2.7%$ and $10.1 pm 9.4%$ for high and low-dimensional latent representations, respectively. For ViTs, high-dimensionality benefits both pre-trained and dataset-optimised SSL. Although location-regularisation improves SSL performance compared to standard SSL methods, pre-trained ViTs show strong generalisation, matching the best-performing location-regularised SSL with F1-scores of $0.795 pm 0.075$ and $0.795 pm 0.077$, respectively. The findings highlight the value of location metadata for SSL regularisation, particularly when using low-dimensional latent representations, and demonstrate strong generalisation of high-dimensional ViTs for seafloor image analysis.

Problem

• Evaluating location-based regularization in self-supervised learning for seafloor imagery
• Assessing impact on CNN and ViT models across diverse seafloor datasets
• Improving downstream classification performance through location metadata integration

Innovation

• Location-regularised self-supervised learning for seafloor imagery
• Evaluating six SSL frameworks with CNN and ViT models
• Location metadata enhances classification performance consistently