MIAYOLO: Multi-Expert and Intra-Class Aggregation Assisted Suburban Building Detection in Unmanned Aerial Vehicle Imagery

Dong Ren,Gan Zhao,Hang Sun,Shun Ren,Li Liu

DOI: https://doi.org/10.1109/lgrs.2024.3367548

IF: 5.343

2024-01-01

IEEE Geoscience and Remote Sensing Letters

Abstract:In recent years, the utilization of unmanned aerial vehicle (UAV) for building detection has become increasingly important in monitoring illegal construction. However, the UAV imagery contains a significantly larger number of completed buildings compared to under construction buildings, which leads to the model exhibiting a bias toward the recognition of completed buildings during the training. Moreover, numerous under construction buildings visually resemble completed buildings, posing a challenge for existing object detection models in distinguishing between these two categories. To address these issues, in this study, we propose multi-expert and intra-class aggregation assisted detection (MIAYOLO) for building detection in UAV images. Specifically, a multi-expert subnetwork is proposed to improve the learning of distinct feature representations for each category, which mitigates the model bias toward the dominant category. Furthermore, an intra-class loss is introduced to aggregate features from the same category while separating instances from the decision boundary. Additionally, the multi-expert subnetwork is discarded to avoid introducing additional computation in the inference stage. Experimental results on our proposed dataset demonstrate that MIAYOLO outperforms state-of-the-art object detection methods, achieving a significant improvement of +4.7 (in terms of mAP50). The code and dataset can available at https://github.com/Game-zhao/MIAYOLO.

imaging science & photographic technology,remote sensing,engineering, electrical & electronic,geochemistry & geophysics