Learning Calibrated-Guidance for Object Detection in Aerial Images

Abstract

Object detection is one of the most fundamental yet challenging research topics in the domain of computer vision. Recently, the study on this topic in aerial images has made tremendous progress. However, complex background and worse imaging quality are obvious problems in aerial object detection. Most state-of-the-art approaches tend to develop elaborate attention mechanisms for the space-time feature calibrations with arduous computational complexity, while surprisingly ignoring the importance of feature calibrations in channel-wise. In this work, we propose a simple yet effective Calibrated-Guidance (CG) scheme to enhance channel communications in a feature transformer fashion, which can adaptively determine the calibration weights for each channel based on the global feature affinity correlations. Specifically, for a given set of feature maps, CG first computes the feature similarity between each channel and the remaining channels as the intermediary calibration guidance. Then, re-representing each channel by aggregating all the channels weighted together via the guidance operation. Our CG is a general module that can be plugged into any deep neural networks, which is named as CG-Net. To demonstrate its effectiveness and efficiency, extensive experiments are carried out on both oriented object detection task and horizontal object detection task in aerial images. Experimental results on two challenging benchmarks (i.e., DOTA and HRSC2016) demonstrate that our CG-Net can achieve the new state-of-the-art performance in accuracy with a fair computational overhead. The source code has been open sourced at https://github.com/WeiZongqi/CG-Net

Shared-Weight-Based Multi-Dimensional Feature Alignment Network for Oriented Object Detection in Remote Sensing Imagery

“…Another line of research focuses on feature alignment in the semantic dimension. CGNet [ 39 ] uses self-attention to enhance communication between pyramid levels. AFF-Det [ 14 , 18 ] maps ROIs to all levels and applies a unified supervisory signal to alleviate the semantic gap.…”

Algorithm for Detecting Small Targets Based on Upgraded YOLOv5s

“…As a result of the tiny scale of small targets, the prediction of centroids is useful for localizing small targets and is a viable approach for small target identification. Transformer was used by WEI et al [3] to detect small targets, and they also put forward the CG-Net (Calibrated-Guidance) idea, which improves the connectivity between channels using the Transformer function. Extracting rich contexts from the many windows that were generated would help train improved target representation.…”

A Tiny Object Detection Approach for Maize Cleaning Operations

“…Therefore, tiny object detection has become one of the most challenging tasks in computer vision [32]. In allusion to the smaller size and higher density of the objects in the aerial images, Wei et al proposed an efficacious calibrated-guidance (CG) scheme to intensify the channel communication through the feature transformer fashion, which could adaptively determine the calibration weights for each channel based on the global feature affinity correlations [33]. The concept of fusion factor was proposed by Gong et al to control the information that delivered from deep layers to the shallow ones, which adapted the feather pyramid network (FPN) to tiny object detection, and its effective value was estimated based on a statistical method [34].…”