Are All Marine Species Created Equal? Performance Disparities in Underwater Object Detection
This work addresses the problem of uneven detection accuracy for marine species in underwater monitoring, which is incremental as it analyzes existing datasets and methods to identify specific bottlenecks.
The paper investigates performance disparities in underwater object detection for marine species, finding that foreground-background discrimination in localization is a key bottleneck and that classification gaps persist even with balanced data, leading to recommendations for dataset balancing based on precision or recall priorities.
Underwater object detection is critical for monitoring marine ecosystems but poses unique challenges, including degraded image quality, imbalanced class distribution, and distinct visual characteristics. Not every species is detected equally well, yet underlying causes remain unclear. We address two key research questions: 1) What factors beyond data quantity drive class-specific performance disparities? 2) How can we systematically improve detection of under-performing marine species? We manipulate the DUO and RUOD datasets to separate the object detection task into localization and classification and investigate the under-performance of the scallop class. Localization analysis using YOLO11 and TIDE finds that foreground-background discrimination is the most problematic stage regardless of data quantity. Classification experiments reveal persistent precision gaps even with balanced data, indicating intrinsic feature-based challenges beyond data scarcity and inter-class dependencies. We recommend imbalanced distributions when prioritizing precision, and balanced distributions when prioritizing recall. Improving under-performing classes should focus on algorithmic advances, especially within localization modules. We publicly release our code and datasets.