The paper introduces an enhanced automated quality assessment network (IBS-AQSNet), a new method for assessing the quality of interactive building segmentation in high-resolution remote sensing imagery. The system combines a robust, pre-trained backbone with a lightweight counterpart for comprehensive feature extraction from imagery and segmentation results. The features are then fused through a simple combination of concatenation, convolution layers, and residual connections. The method also includes a multi-scale differential quality assessment decoder, capable of pinpointing areas where segmentation result is either missed or mistaken. This method, tested on a dataset of over 39,198 buildings, has set a new standard in the field of automating segmentation quality assessment.
Publication date: 19 Jan 2024
Project Page: Not Provided
Paper: https://arxiv.org/pdf/2401.09828
