AutoPET III Challenge: PET/CT Semantic Segmentation
This work addresses the problem of precise lesion segmentation in medical imaging for clinicians, but it appears incremental as it combines existing methods without introducing new paradigms.
The study tackled lesion segmentation in PET/CT images for the AutoPET III challenge by implementing a two-stage deep learning approach with coarse segmentation using DynUNet and refinement via an ensemble of SwinUNETR, SegResNet, and UNet models, but no concrete performance numbers were provided.
In this study, we implemented a two-stage deep learning-based approach to segment lesions in PET/CT images for the AutoPET III challenge. The first stage utilized a DynUNet model for coarse segmentation, identifying broad regions of interest. The second stage refined this segmentation using an ensemble of SwinUNETR, SegResNet, and UNet models. Preprocessing involved resampling images to a common resolution and normalization, while data augmentation techniques such as affine transformations and intensity adjustments were applied to enhance model generalization. The dataset was split into 80% training and 20% validation, excluding healthy cases. This method leverages multi-stage segmentation and model ensembling to achieve precise lesion segmentation, aiming to improve robustness and overall performance.