Foreign object segmentation in chest x-rays through anatomy-guided shape insertion
This addresses the problem of insufficient annotated data for foreign object segmentation in chest x-rays, particularly for postoperative follow-ups, with an incremental approach that reduces annotation effort.
The paper tackles the challenge of instance segmentation for foreign objects in chest radiographs by generating synthetic data through anatomy-guided shape insertion and cut-paste augmentations, achieving performance comparable to fully supervised models while using 93% fewer manual annotations.
In this paper, we tackle the challenge of instance segmentation for foreign objects in chest radiographs, commonly seen in postoperative follow-ups with stents, pacemakers, or ingested objects in children. The diversity of foreign objects complicates dense annotation, as shown in insufficient existing datasets. To address this, we propose the simple generation of synthetic data through (1) insertion of arbitrary shapes (lines, polygons, ellipses) with varying contrasts and opacities, and (2) cut-paste augmentations from a small set of semi-automatically extracted labels. These insertions are guided by anatomy labels to ensure realistic placements, such as stents appearing only in relevant vessels. Our approach enables networks to segment complex structures with minimal manually labeled data. Notably, it achieves performance comparable to fully supervised models while using 93\% fewer manual annotations.