Dual-Stream Spectral Decoupling Distillation for Remote Sensing Object Detection

TL;DR

This paper introduces DS2D2, a spectral decoupling distillation method that improves remote sensing object detection by addressing feature entanglement and subtle discrepancies between teacher and student models.

Contribution

The paper proposes a novel spectral decomposition-based distillation architecture that effectively preserves spatial features and extracts implicit knowledge for remote sensing object detection.

Findings

Achieves 4.2% AP50 improvement on DIOR with RetinaNet

Outperforms existing distillation methods on DIOR and DOTA datasets

Validates effectiveness through extensive experiments

Abstract

Knowledge distillation is an effective and hardware-friendly method, which plays a key role in lightweighting remote sensing object detection. However, existing distillation methods often encounter the issue of mixed features in remote sensing images (RSIs), and neglect the discrepancies caused by subtle feature variations, leading to entangled knowledge confusion. To address these challenges, we propose an architecture-agnostic distillation method named Dual-Stream Spectral Decoupling Distillation (DS2D2) for universal remote sensing object detection tasks. Specifically, DS2D2 integrates explicit and implicit distillation grounded in spectral decomposition. Firstly, the first-order wavelet transform is applied for spectral decomposition to preserve the critical spatial characteristics of RSIs. Leveraging this spatial preservation, a Density-Independent Scale Weight (DISW) is designed to address the challenges of dense and small object detection common in RSIs. Secondly, we show implicit knowledge hidden in subtle student-teacher feature discrepancies, which significantly influence predictions when activated by detection heads. This implicit knowledge is extracted via full-frequency and high-frequency amplifiers, which map feature differences to prediction deviations. Extensive experiments on DIOR and DOTA datasets validate the effectiveness of the proposed method. Specifically, on DIOR dataset, DS2D2 achieves improvements of 4.2% in AP50 for RetinaNet and 3.8% in AP50 for Faster R-CNN, outperforming existing distillation approaches. The source code will be available at https://github.com/PolarAid/DS2D2.