New DyABD benchmark dataset advances abdominal muscle segmentation in dynamic MRI

Researchers have introduced DyABD, a new benchmark dataset for segmenting abdominal muscles in dynamic MRI scans. This dataset is unique as it captures MRIs of patients performing exercises, leading to significant anatomical variability and making it a challenging benchmark for existing segmentation models. The evaluation revealed that current techniques achieve an average Dice Coefficient of 0.82, indicating substantial room for improvement in medical image segmentation. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Establishes a new, challenging benchmark for medical image segmentation, potentially driving advancements in AI-powered diagnostic tools.

RANK_REASON The cluster describes the introduction of a new benchmark dataset and evaluation of existing models in a specific research domain.

Read on arXiv cs.CV →

• DyABD
• MRI
• Dice Coefficient

• paper
• other