New AQUA20 dataset targets challenging underwater species classification

Researchers have introduced AQUA20, a new benchmark dataset designed to improve underwater species classification. The dataset contains 8,171 images of 20 marine species, specifically curated to address challenges like turbidity, low illumination, and occlusion. Experiments were conducted on thirteen deep learning models, with ConvNeXt demonstrating the highest accuracy at 90.69% Top-1 and 88.92% F1-score, though other models showed varying performance trade-offs. The study also includes an explainability analysis using GRAD-CAM and LIME to interpret model behavior. AI

IMPACT Provides a new resource for advancing computer vision models in challenging underwater environments.

RANK_REASON Publication of a new benchmark dataset and accompanying research paper. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• AQUA20
• ConvNeXt
• GRAD-CAM
• MobileNetV2
• SqueezeNet
• Taufikur Rahman Fuad
• ViT

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →