Modeling annotator preference and stochastic annotation error for medical image segmentation

Zehui Liao, Shishuai Hu, Yutong Xie, Yong Xia

Abstract: Highlights•We propose PADL to address annotator-related biases in medical image segmentation.•We handle that biases as a mix of annotator preference and stochastic errors in PADL.•Our PADL outperforms other methods on two medical image segmentation benchmarks.