Revisiting Superpixels for Active Learning in Semantic Segmentation With Realistic Annotation Costs

State-of-the-art methods for semantic segmentation are based on deep neural networks that are known to be datahungry. Region-based active learning has shown to be a promising method for reducing data annotation costs. A key design choice for region-based AL is whether to use regularly-shaped regions (e.g., rectangles) or irregularlyshaped region (e.g., superpixels). In this work, we address this question under realistic, click-based measurement of annotation costs. In particular, we revisit the use of superpixels and demonstrate that the inappropriate choice of cost measure (e.g., the percentage of labeled pixels), may cause the effectiveness of the superpixel-based approach to be under-estimated. We benchmark the superpixel-based approach against the traditional "rectangle+polygon"-based approach with annotation cost measured in clicks, and show that the former outperforms on both Cityscapes and PAS-CAL VOC. We further propose a class-balanced acquisition function to boost the performance of the superpixel-based approach and demonstrate its effectiveness on the evaluation datasets. Our results strongly argue for the use of superpixel-based AL for semantic segmentation and highlight the importance of using realistic annotation costs in evaluating such methods.