A Hybrid Transformer-Convolutional Neural Network for Segmentation of Intracerebral Hemorrhage and Perihematomal Edema on Non-Contrast Head Computed Tomography (CT) with Uncertainty Quantification to Improve Confidence DOI Creative Commons
Tran Anh Tuan, Dmitriy Desser, Tal Zeevi

et al.

Bioengineering, Journal Year: 2024, Volume and Issue: 11(12), P. 1274 - 1274

Published: Dec. 15, 2024

Intracerebral hemorrhage (ICH) and perihematomal edema (PHE) are key imaging markers of primary secondary brain injury in hemorrhagic stroke. Accurate segmentation quantification ICH PHE can help with prognostication guide treatment planning. In this study, we combined Swin-Unet Transformers nnU-NETv2 convolutional network for on non-contrast head CTs. We also applied test-time data augmentations to assess individual-level prediction uncertainty, ensuring high confidence prediction. The model was trained 1782 CT scans from a multicentric trial tested two independent datasets Yale (n = 396) University Berlin Charité Hospital Medical Center Hamburg-Eppendorf 943). Model performance evaluated the Dice coefficient Volume Similarity (VS). Our dual Swin-nnUNET achieved median (95% interval) 0.93 (0.90–0.95) VS 0.97 (0.95–0.98) ICH, 0.70 (0.64–0.75) 0.87 (0.80–0.93) cohort. 0.86 (0.80–0.90) 0.91 (0.85–0.95) 0.65 (0.56–0.70) (0.77–0.93) Berlin/Hamburg-Eppendorf Prediction uncertainty associated lower accuracy, smaller ICH/PHE volumes, infratentorial location. results highlight benefits transformer-convolutional neural architecture augmentation quantification.

Language: Английский

AATCT-IDS: A benchmark Abdominal Adipose Tissue CT Image Dataset for image denoising, semantic segmentation, and radiomics evaluation DOI

Zhiyu Ma,

Chen Li, Tianming Du

et al.

Computers in Biology and Medicine, Journal Year: 2024, Volume and Issue: 177, P. 108628 - 108628

Published: May 21, 2024

Language: Английский

Citations

1
A HybridOpt approach for early Alzheimer’s Disease diagnostics with Ant Lion Optimizer (ALO) DOI Creative Commons

A. Sasithradevi,

Chanthini Baskar,

S. Shoba

et al.

Alexandria Engineering Journal, Journal Year: 2024, Volume and Issue: 109, P. 112 - 125

Published: Sept. 4, 2024

Language: Английский

Citations

0
RPE-Diff: A Relative Position Encoding Diffusion Model for Perirenal Fat Segmentation in Metabolic Syndrome DOI
Shuai Ye, Tianming Du, Frank Kulwa

et al.

Lecture notes in computer science, Journal Year: 2024, Volume and Issue: unknown, P. 155 - 170

Published: Dec. 12, 2024

Language: Английский

Citations

0
A Hybrid Transformer-Convolutional Neural Network for Segmentation of Intracerebral Hemorrhage and Perihematomal Edema on Non-Contrast Head Computed Tomography (CT) with Uncertainty Quantification to Improve Confidence DOI Creative Commons
Tran Anh Tuan, Dmitriy Desser, Tal Zeevi

et al.

Bioengineering, Journal Year: 2024, Volume and Issue: 11(12), P. 1274 - 1274

Published: Dec. 15, 2024

Intracerebral hemorrhage (ICH) and perihematomal edema (PHE) are key imaging markers of primary secondary brain injury in hemorrhagic stroke. Accurate segmentation quantification ICH PHE can help with prognostication guide treatment planning. In this study, we combined Swin-Unet Transformers nnU-NETv2 convolutional network for on non-contrast head CTs. We also applied test-time data augmentations to assess individual-level prediction uncertainty, ensuring high confidence prediction. The model was trained 1782 CT scans from a multicentric trial tested two independent datasets Yale (n = 396) University Berlin Charité Hospital Medical Center Hamburg-Eppendorf 943). Model performance evaluated the Dice coefficient Volume Similarity (VS). Our dual Swin-nnUNET achieved median (95% interval) 0.93 (0.90–0.95) VS 0.97 (0.95–0.98) ICH, 0.70 (0.64–0.75) 0.87 (0.80–0.93) cohort. 0.86 (0.80–0.90) 0.91 (0.85–0.95) 0.65 (0.56–0.70) (0.77–0.93) Berlin/Hamburg-Eppendorf Prediction uncertainty associated lower accuracy, smaller ICH/PHE volumes, infratentorial location. results highlight benefits transformer-convolutional neural architecture augmentation quantification.

Language: Английский

Citations

0