# Label-efficient underwater image classification with logistic regression on frozen foundation model embeddings

arXiv: https://arxiv.org/abs/2604.00313 · HTML: https://mlsanity.com/paper/2604.00313/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 22 of its 34 references that are in the map.
- First posted on arXiv 31 Mar 2026.

## Builds on (most cited first, 22 of 22)

- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Emerging Properties in Self-Supervised Vision Transformers](https://mlsanity.com/paper/2104.14294/) (2021)
- [You Only Look Once: Unified, Real-Time Object Detection](https://mlsanity.com/paper/1506.02640/) (2015)
- [SAM 3: Segment Anything with Concepts](https://mlsanity.com/paper/2511.16719/) (2025)
- [Vision Transformers Need Registers](https://mlsanity.com/paper/2309.16588/) (2023)
- [Fast Underwater Image Enhancement for Improved Visual Perception](https://mlsanity.com/paper/1903.09766/) (2019)
- [A Comprehensive Review of YOLO Architectures in Computer Vision: From YOLOv1 to YOLOv8 and YOLO-NAS](https://mlsanity.com/paper/2304.00501/) (2023)
- [The Shape of Learning Curves: a Review](https://mlsanity.com/paper/2103.10948/) (2021)
- [Do We Train on Test Data? Purging CIFAR of Near-Duplicates](https://mlsanity.com/paper/1902.00423/) (2019)
- [FathomNet: A global image database for enabling artificial intelligence in the ocean](https://mlsanity.com/paper/2109.14646/) (2021)
- [Computer Vision and Deep Learning for Fish Classification in Underwater Habitats: A Survey](https://mlsanity.com/paper/2203.06951/) (2022)
- [Fish-Vista: A Multi-Purpose Dataset for Understanding & Identification of Traits from Images](https://mlsanity.com/paper/2407.08027/) (2024)
- [AQUA20: A Benchmark Dataset for Underwater Species Classification under Challenging Conditions](https://mlsanity.com/paper/2506.17455/) (2025)
- [Empowering DINO Representations for Underwater Instance Segmentation via Aligner and Prompter](https://mlsanity.com/paper/2511.08334/) (2025)
- [BenthicNet: A global compilation of seafloor images for deep learning applications](https://mlsanity.com/paper/2405.05241/) (2024)
- [Multi-Label Plant Species Classification with Self-Supervised Vision Transformers](https://mlsanity.com/paper/2407.06298/) (2024)
- [Unlocking the potential of deep learning for marine ecology: overview, applications, and outlook](https://mlsanity.com/paper/2109.14737/) (2021)
- [Vision Transformers for Zero-Shot Clustering of Animal Images: A Comparative Benchmarking Study](https://mlsanity.com/paper/2602.03894/) (2026)
- [Mitigating Domain Drift in Multi Species Segmentation with DINOv2: A Cross-Domain Evaluation in Herbicide Research Trials](https://mlsanity.com/paper/2508.07514/) (2025)
- [Linear probing enables Ship-Radiated Noise recognition with pretrained audio embeddings](https://mlsanity.com/paper/2601.08358/) (2026)
- [Multi-label classification for multi-temporal, multi-spatial coral reef condition monitoring using vision foundation model with adapter learning](https://mlsanity.com/paper/2503.23012/) (2025)

## Abstract (arXiv)

Underwater image classification is constrained by the cost of annotation and by the computational and methodological requirements of task-specific model development. We investigate whether frozen general-purpose foundation-model embeddings can reduce these requirements by extracting DINOv3 ViT-B/16 embeddings and training only a logistic regression classifier on the AQUA20 benchmark. We evaluate the approach across a range of annotation budgets, a repeated 80% training-subsample evaluation, and a full-training refit. With only 13 labelled images per category, corresponding to approximately 4% of the benchmark's official training partition, mean macro F1 reaches 81.8%; with 144 images per category it reaches 88.5%, compared to the published fully supervised ConvNeXt point estimate of 88.9% obtained with the complete training set (benchmark results reported without run-to-run variability). Using all official training labels, macro F1 reaches 91.5% (bootstrap 95% CI: 89.0-93.7%). Sensitivity analyses show that the main findings remain stable across ordinary downstream implementation choices, and persist after removing duplicate and near-duplicate test images identified in an audit of the official split. Preliminary evaluation on a second dataset suggests that overall performance level and the shape of the label efficiency curve are not unique to the AQUA20 dataset. Because the DINOv3 backbone remains frozen and only the downstream classifier is fitted, the approach avoids task-specific neural-network training and can be executed on commodity hardware. These findings establish linear classification on frozen foundation-model embeddings as a practical baseline for label-efficient underwater image classification.