From Whole-Body to Abdomen: Streamlined Segmentation of Organs and Tumors via Semi-Supervised Learning and Efficient Coarse-to-Fine Inference

Shoujin Huang, Huaishui Yang, Lifeng Mei, Tan Zhang, Shaojun Liu, Mengye Lyu

Keywords: FLARE2023 ｜ Segmentation

TL;DR: We address challenges in medical image analysis for abdominal organs and tumors. We use advanced techniques, including pseudo-labels, and a new segmentation approach to significantly speed up the process with high accuracy.

Abstract: Precise and automated segmentation of abdominal organs and tumors is an important research area of medical image analysis. This domain faces three key challenges: the presence of partially labeled training data that can mislead model training, the variable morphologies of tumors complicating the segmentation process, and the computationally demanding nature of inference in whole/half-body CT scans. In our study, we leverage advanced techniques to generate pseudo-labels, thereby adequately addressing the limitations of partially annotated datasets in a semi-supervised manner. Furthermore, we introduce a novel perspective that allows the segmentation of whole/half-body CT scans to be streamlined into focused abdominal segmentation. To achieve this, we re-engineered the nnU-Net V2 inference engine to incorporate a coarse-to-fine strategy, leading to a remarkable 15x speed-up by eliminating extraneous regions. Our approach yields a mean Dice Similarity Coefficient (DSC) of 90.75/47.95 and a Normalized Surface Dice (NSD) of 95.54/40.16 for organ and tumor segmentation, respectively, in the FLARE 2023 validation dataset. Importantly, our method accomplishes these results with an average processing time of only 27.47 seconds per case.

Submission Number: 6