Boundary Loss for Remote Sensing Imagery Semantic Segmentation

Abstract

In response to the growing importance of geospatial data, its analysis including semantic segmentation becomes an increasingly popular task in computer vision today. Convolutional neural networks are powerful visual models that yield hierarchies of features and practitioners widely use them to process remote sensing data. When performing remote sensing image segmentation, multiple instances of one class with precisely defined boundaries are often the case, and it is crucial to extract those boundaries accurately. The accuracy of segments boundaries delineation influences the quality of the whole segmented areas explicitly. However, widely-used segmentation loss functions such as BCE, IoU loss or Dice loss do not penalize misalignment of boundaries sufficiently. In this paper, we propose a novel loss function, namely a differentiable surrogate of a metric accounting accuracy of boundary detection. We can use the loss function with any neural network for binary segmentation. We performed validation of our loss function with various modifications of UNet on a synthetic dataset, as well as using real-world data (ISPRS Potsdam, INRIA AIL). Trained with the proposed loss function, models outperform baseline methods in terms of IoU score.

PiCoCo: Pixelwise Contrast and Consistency Learning for Semisupervised Building Footprint Segmentation

“…Wu et al regularized the region-based CE loss based on the boundary loss for extracting building segments and outlines [9]. Bokhovkin et al proposed a surrogate loss to penalize the misalignment of building boundaries and achieved the out-performance than the commonly utilized CE and Dice losses [42]. Although the above-mentioned methods have significantly improved the performance of the building footprint generation, few work have been investigated to learn the segmentation models based on limited RS images with annotations.…”

Segmentation of Pancreatic Ductal Adenocarcinoma (PDAC) and surrounding vessels in CT images using deep convolutional neural networks and texture descriptors

“…Because of low contrast around the boundary of pancreas and tumor, exact determination of tumor and healthy tissue boundary can be challenging. To diminish this issue, we also use the differentiable version of the boundary loss proposed in (41). The boundaries of the ground truth and predicted segments are obtained using max-pooling operation as follow: where a pixel-wise max-pooling operation is applied to obtain the inverted ground truth and predicted binary segments with a kernel size θ 0 .…”

Semantic Segmentation of Aerial Imagery via Split-Attention Networks with Disentangled Nonlocal and Edge Supervision

“…Considering the imbalance of the number of edge pixels and non-edge pixels in the image, we draw inspiration from the boundary loss [39] and use the following function as the loss function for edge detection:…”