Oriented Tiny Object Detection: A Dataset, Benchmark, and Dynamic Unbiased Learning
This work addresses the under-explored challenge of oriented tiny object detection, which is prevalent in real-world applications like aerial imagery, by providing a comprehensive dataset and method to reduce biases in detection pipelines.
The paper tackles the problem of detecting oriented tiny objects by introducing a new dataset (AI-TOD-R) with the smallest object sizes, a benchmark, and a dynamic coarse-to-fine learning scheme (DCFL) that mitigates learning biases, achieving state-of-the-art accuracy across eight datasets.
Detecting oriented tiny objects, which are limited in appearance information yet prevalent in real-world applications, remains an intricate and under-explored problem. To address this, we systemically introduce a new dataset, benchmark, and a dynamic coarse-to-fine learning scheme in this study. Our proposed dataset, AI-TOD-R, features the smallest object sizes among all oriented object detection datasets. Based on AI-TOD-R, we present a benchmark spanning a broad range of detection paradigms, including both fully-supervised and label-efficient approaches. Through investigation, we identify a learning bias presents across various learning pipelines: confident objects become increasingly confident, while vulnerable oriented tiny objects are further marginalized, hindering their detection performance. To mitigate this issue, we propose a Dynamic Coarse-to-Fine Learning (DCFL) scheme to achieve unbiased learning. DCFL dynamically updates prior positions to better align with the limited areas of oriented tiny objects, and it assigns samples in a way that balances both quantity and quality across different object shapes, thus mitigating biases in prior settings and sample selection. Extensive experiments across eight challenging object detection datasets demonstrate that DCFL achieves state-of-the-art accuracy, high efficiency, and remarkable versatility. The dataset, benchmark, and code are available at https://chasel-tsui.github.io/AI-TOD-R/.