Scholarlink

<p>Robust visual recognition remains a significant challenge in underwater environments due to complex distortions such as turbidity, low illumination, and occlusion, which severely degrade the performance of standard vision systems. This study introduces AQUA20, a dataset comprising 8,171 images of 20 ecologically significant marine species. AQUA20 provides greater multi-species diversity and real-world variability than other public datasets, such as Fish4Knowledge and WildFish, which are limited in terms of taxonomic scope or environmental complexity. We benchmarked this dataset by evaluating models ranging from lightweight CNNs suitable for edge deployment (e.g., SqueezeNet and MobileNetV2) to modern vision transformers (e.g., ViT and ConvNeXt) for marine species classification under challenging conditions. ConvNeXt achieved the best performance, with a Top-3 accuracy of 98.82% and Top-1 accuracy of 90.69%, as well as the highest overall F1-score of 88.92%. The results include the trade-offs between complexity and performance. We also provide an extensive explainability analysis using GRAD-CAM and LIME to interpret the strengths and pitfalls of the models, which further reveals that high-performing models focus on biologically relevant features, whereas lower-performing models are distracted by background noise, providing a clear pathway for future improvements in the model architectures. The dataset is publicly available at: <a href="https://huggingface.co/datasets/taufiktrf/AQUA20">https://huggingface.co/datasets/taufiktrf/AQUA20</a>.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

AQUA20: A Benchmark Dataset for Underwater Species Classification under Challenging Conditions

  • 期刊论文

  • 发表日期：

    2026年3月27日

查看全文
  • Taufikur Rahman Fuad,
  • Sabbir Ahmed,
  • Shahriar Ivan

摘要

Robust visual recognition remains a significant challenge in underwater environments due to complex distortions such as turbidity, low illumination, and occlusion, which severely degrade the performance of standard vision systems. This study introduces AQUA20, a dataset comprising 8,171 images of 20 ecologically significant marine species. AQUA20 provides greater multi-species diversity and real-world variability than other public datasets, such as Fish4Knowledge and WildFish, which are limited in terms of taxonomic scope or environmental complexity. We benchmarked this dataset by evaluating models ranging from lightweight CNNs suitable for edge deployment (e.g., SqueezeNet and MobileNetV2) to modern vision transformers (e.g., ViT and ConvNeXt) for marine species classification under challenging conditions. ConvNeXt achieved the best performance, with a Top-3 accuracy of 98.82% and Top-1 accuracy of 90.69%, as well as the highest overall F1-score of 88.92%. The results include the trade-offs between complexity and performance. We also provide an extensive explainability analysis using GRAD-CAM and LIME to interpret the strengths and pitfalls of the models, which further reveals that high-performing models focus on biologically relevant features, whereas lower-performing models are distracted by background noise, providing a clear pathway for future improvements in the model architectures. The dataset is publicly available at: https://huggingface.co/datasets/taufiktrf/AQUA20.