Patch Ensembles for Robust Salmon Re-Identification with Weak Trajectory Labels

2026-05-18

AI Summary

This work addresses the challenges of large population scale, high annotation cost, and bias introduced by weak trajectory labels in salmon re-identification within commercial aquaculture pens. The authors propose a patch-based re-identification framework that leverages side-line localization to guide texture anchor positioning, integrates multi-patch features for identity discrimination, and introduces a cross-camera test set to evaluate generalization. By innovatively incorporating a side-line-guided texture anchoring mechanism and a multi-patch ensemble strategy, the method effectively mitigates label-induced bias. Experimental results demonstrate significant performance gains: same-trajectory validation mAP improves from 0.932 to 0.965, while cross-camera mAP rises markedly from 0.609 to 0.860, confirming the approach’s robustness and superior generalization capability in complex real-world scenarios.

Abstract

Salmon re-identification in commercial net-pens is challenging due to large populations, which impose strict accuracy requirements and make large-scale labeled data acquisition infeasible. Trajectory IDs can be used as proxy labels, but this introduces trajectory-ID bias. To address these challenges, we propose a patch-based re-identification framework that fuses patch-level predictions into a salmon identity decision. A key component is the prediction of the salmon's lateral line, enabling extraction of texture-anchored patches and patch slices. To enable realistic evaluation, we introduce an experimental setup using multiple cameras placed 6 m apart, allowing the same fish to be recorded in different trajectories. This enables the construction of a cross-camera test set through manual match confirmation. Our ensemble approach outperforms the full-image baseline in same-trajectory validation (0.932 to 0.965 mAP) and cross-camera testing (0.609 to 0.860 mAP). The substantial improvements in the cross-camera setting demonstrate improved generalizability and robustness. Code and data: https://github.com/espenbh/salmon-reid-patch-ensemble.

Problem

• salmon re-identification
• weak trajectory labels
• trajectory-ID bias
• large-scale identification
• cross-camera generalization

Innovation

• patch ensemble
• salmon re-identification
• lateral line prediction
• weak trajectory labels
• cross-camera evaluation