Sorry, you need to enable JavaScript to visit this website.

Extracting and Distilling Direction-adaptive Knowledge for Lightweight Object Ddetection in Remote Sensing Images

Citation Author(s):
Zhanchao Huang, Wei Li, Ran Tao
Submitted by:
Zhanchao Huang
Last updated:
10 May 2022 - 12:10am
Document Type:
Zhanchao Huang
Paper Code:


Recently, some lightweight convolutional neural network (CNN) models have been proposed for airborne or spaceborne remote sensing object detection (RSOD) tasks. However, these lightweight detectors suffer from performance degradation due to the compromise of limited computing resources on embedded devices. In order to narrow this performance gap, a direction-adaptive knowledge extraction and distillation (DKED) method is proposed. Specifically, a dynamic directional convolution (DDC) is developed to extract the typical arbitrary-oriented features, and a direction-adaptive knowledge distillation (DKD) strategy is designed for guiding the lightweight model to learn the intrinsic knowledge of the RSOD task from the high-performance model. Experiments on public datasets demonstrate that the proposed method can effectively improve the performance of the lightweight RSOD model without additional inference costs.

0 users have voted: