Reliable Label-Supervised Pixel Attention Mechanism for Weakly Supervised Building Segmentation in UAV Imagery

Abstract

Building segmentation for Unmanned Aerial Vehicle (UAV) imagery usually requires pixel-level labels, which are time-consuming and expensive to collect. Weakly supervised semantic segmentation methods for image-level labeling have recently achieved promising performance in natural scenes, but there have been few studies on UAV remote sensing imagery. In this paper, we propose a reliable label-supervised pixel attention mechanism for building segmentation in UAV imagery. Our method is based on the class activation map. However, classification networks tend to capture discriminative parts of the object and are insensitive to over-activation; therefore, class activation maps cannot directly guide segmentation network training. To overcome these challenges, we first design a Pixel Attention Module that captures rich contextual relationships, which can further mine more discriminative regions, in order to obtain a modified class activation map. Then, we use the initial seeds generated by the classification network to synthesize reliable labels. Finally, we design a reliable label loss, which is defined as the sum of the pixel-level differences between the reliable labels and the modified class activation map. Notably, the reliable label loss can handle over-activation. The preceding steps can significantly improve the quality of the pseudo-labels. Experiments on our home-made UAV data set indicate that our method can achieve 88.8% mIoU on the test set, outperforming previous state-of-the-art weakly supervised methods.

Beyond Supervised Learning in Remote Sensing: A Systematic Review of Deep Learning Approaches

“…In fact, preparing reliable and accurate labels for large-scale multi-label image classification is a difficult and expensive process that also involves human errors. To reduce the effect of labeling uncertainties WSL methods have been investigated in various fields of classification tasks, such as image-level annotated WSL for building segmentation using VHR RGB imagery [80], and tree species classification using VHR multispectral and LiDAR [47], and point-level annotated WSL for road network [46] and water body [81] segmentations. Furthermore, as shown by , even weak training samples in pointlevel or image-level annotations combined with a deep model such as U-Net can outperform supervised baseline methods such as support vector machine (SVM) [82].…”

Dual-Channel Semi-Supervised Adversarial Network for Building Segmentation from UAV-Captured Images

“…In the field of UAV images, as highlighted in the literature [58], the distance between the building and the camera may lead to reduced image clarity, and the size of the building will also show significant changes. As shown in Figure 2 of reference [59], relying only on an interpolation algorithm to compensate sharpness leads to instability in texture details and spatial information, seriously affecting the accuracy of extraction.…”