From Spatial to Spectral: An Efficient, Frequency-Guided Feature Representation Learner for Small Object Detection
For practitioners in small object detection (e.g., drone, satellite, pedestrian detection), this work offers a parameter-efficient alternative to spatial-only detectors by leveraging frequency-domain cues.
Small object detection suffers from feature scarcity due to spatial-domain operations discarding high-frequency details. The authors propose a frequency-guided feature representation framework (DERNet) that shifts processing from spatial to spectral domain, achieving consistent gains across multiple benchmarks (e.g., VisDrone2019, UAVDT, TinyPerson, DOTAv1) and outperforming YOLOv11 with only 1/6 of the parameters.
Efficient small object detection is bottlenecked by the inherent feature scarcity of tiny targets, which is further aggravated by operations of spatial-domain detectors that indiscriminately discard critical high-frequency details. Recovering these fragile cues within the spatial domain is notoriously difficult, as it often requires computationally expensive architectural upscaling that inadvertently amplifies background noise. To bridge this gap, we propose a paradigm \textbf{shift from spatial to spectral} feature processing, introducing a holistic solution with the following novelty: (1) A versatile \textbf{Frequency-Guided Feature Representation framework} that generalizes across diverse detector architectures (both CNN and Transformer-based), offering a robust alternative to spatial-only feature extraction; (2) The unified \textbf{Decompose--Enhance--Reconstruct (DER)} operator, instantiated via three \textbf{lightweight, plug-and-play} modules -- Wavelet-Difference Gate (WDG), Log-Gabor Enhancer (LGE), and Frequency-Driven Head (FDHead) -- to systematically inject frequency-aware modulation into the backbone, neck, and head. This mechanism decouples feature modeling from resolution reduction, capturing discriminative high-frequency components to enable accurate localization with significantly reduced parameter redundancy; (3) Extensive validation on multi-domain benchmarks (VisDrone2019, UAVDT, TinyPerson, DOTAv1) demonstrating consistent gains. Notably, our proposed \textbf{DERNet} series outperforms YOLOv11 models under the same scale while requiring \textbf{only 1/6 of the parameters}, backed by rigorous spectral diagnostics and error decomposition analysis.