New backpropagation-free framework for thyroid nodule segmentation

Researchers have developed MedSaab-US, a novel framework for segmenting thyroid nodules in ultrasound images that does not rely on backpropagation or deep learning. This approach combines multi-level Discrete Wavelet Transform with multi-scale Saab transforms to extract features, which are then processed by an XGBoost classifier. MedSaab-US achieves a mean Dice coefficient of 0.4784 on the TN3K dataset, with a small model footprint and CPU-only inference capabilities, offering a potential alternative for resource-constrained environments. AI

IMPACT Offers a potential alternative to deep learning for medical image segmentation in resource-constrained settings.

RANK_REASON The item describes a new research paper detailing a novel framework for medical image segmentation. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• Discrete Wavelet Transform
• MedSaab-US
• Mohammad Amanour Rahman
• Saab transform
• Thyroid Nodule Segmentation
• TN3K dataset
• XGBoost

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →