Sine Wave Normalization for Deep Learning-Based Tumor Segmentation in CT/PET Imaging
This work addresses the problem of accurate tumor segmentation in medical imaging for clinicians, but it appears incremental as it focuses on a specific normalization technique within an existing challenge framework.
The authors tackled automated tumor segmentation in CT/PET scans by introducing a normalization block called SineNormal, which applies sine transformations to PET data to enhance lesion detection, resulting in improved segmentation accuracy for multitracer PET datasets.
This report presents a normalization block for automated tumor segmentation in CT/PET scans, developed for the autoPET III Challenge. The key innovation is the introduction of the SineNormal, which applies periodic sine transformations to PET data to enhance lesion detection. By highlighting intensity variations and producing concentric ring patterns in PET highlighted regions, the model aims to improve segmentation accuracy, particularly for challenging multitracer PET datasets. The code for this project is available on GitHub (https://github.com/BBQtime/Sine-Wave-Normalization-for-Deep-Learning-Based-Tumor-Segmentation-in-CT-PET).