KIKI ‐net: cross‐domain convolutional neural networks for reconstructing undersampled magnetic resonance images

2018 ◽  
Vol 80 (5) ◽  
pp. 2188-2201 ◽  
Author(s):  
Taejoon Eo ◽  
Yohan Jun ◽  
Taeseong Kim ◽  
Jinseong Jang ◽  
Ho‐Joon Lee ◽  
...  
Keyword(s):  
Neural Networks ◽  
Magnetic Resonance ◽  
Convolutional Neural Networks ◽  
Magnetic Resonance Images ◽  
Cross Domain

scholarly journals Comparative Study of Encoder–Decoder-Based Convolutional Neural Networks in Cartilage Delineation from Knee Magnetic Resonance Images

2020 ◽  
Vol 16 ◽  
Author(s):  
Ching Wai Yong ◽  
Khin Wee Lai ◽  
Belinda Pingguan Murphy ◽  
Yan Chai Hum
Keyword(s):  
Neural Networks ◽  
Magnetic Resonance ◽  
Convolutional Neural Networks ◽  
Soft Tissues ◽  
Early Stage ◽  
Joint Inflammation ◽  
Magnetic Resonance Images ◽  
Mr Images ◽  
Daily Lives ◽  
Osteoarthritis Initiative

Background: Osteoarthritis (OA) is a common degenerative joint inflammation which may lead to disability. Although OA is not lethal, this disease will remarkably affect patient’s mobility and their daily lives. Detecting OA at an early stage allows for early intervention and may slow down disease progression. Introduction: Magnetic resonance imaging is a useful technique to visualize soft tissues within the knee joint. Cartilage delineation in magnetic resonance (MR) images helps in understanding the disease progressions. Convolutional neural networks (CNNs) have shown promising results in computer vision tasks, and various encoder–decoder-based segmentation neural networks are introduced in the last few years. However, the performances of such networks are unknown in the context of cartilage delineation. Methods: This study trained and compared 10 encoder–decoder-based CNNs in performing cartilage delineation from knee MR images. The knee MR images are obtained from Osteoarthritis Initiative (OAI). The benchmarking process is to compare various CNNs based on the physical specifications and segmentation performances. Results: LadderNet has the least trainable parameters with model size of 5 MB. UNetVanilla crowned the best performances by having 0.8369, 0.9108, and 0.9097 on JSC, DSC, and MCC. Conclusion: UNetVanilla can be served as a benchmark for cartilage delineation in knee MR images while LadderNet served as alternative if there are hardware limitations during production.

scholarly journals Liver Volumetry from Magnetic Resonance Images with Convolutional Neural Networks

2021 ◽  
Author(s):  
Sara L. Saunders ◽  
Justin M. Clark ◽  
Kyle Rudser ◽  
Anil Chauhan ◽  
Justin R. Ryder ◽  
...  
Keyword(s):  
Neural Networks ◽  
Magnetic Resonance ◽  
Convolutional Neural Networks ◽  
Liver Volume ◽  
Network Models ◽  
Magnetic Resonance Images ◽  
Liver Fat ◽  
Dice Similarity Coefficient ◽  
Neural Network Models ◽  
Mri Scans

AbstractPurposeTo determine which types of magnetic resonance images give the best performance when used to train convolutional neural networks for liver segmentation and volumetry.MethodsAbdominal MRI scans were performed on 42 adolescents with obesity. Scans included Dixon imaging (giving water, fat, and T2* images) and low-resolution T2-weighted anatomical scans. Multiple convolutional neural network models using a 3D U-Net architecture were trained with different input images. Whole-liver manual segmentations were used for reference.Segmentation performance was measured using the Dice similarity coefficient (DSC) and 95% Hausdorff distance. Liver volume accuracy was evaluated using bias, precision, and normalized root mean square error (NRMSE).ResultsThe models trained using both water and fat images performed best, giving DSC = 0.94 and NRMSE = 4.2%. Models trained without the water image as input all performed worse, including in participants with elevated liver fat. Models using the T2-weighted anatomical images underperformed the Dixon-based models, but provided acceptable performance (DSC ≥ 0.92, NMRSE ≤ 6.6%) for use in longitudinal pediatric obesity interventions.ConclusionThe model using Dixon water and fat images as input gave the best performance, with results comparable to inter-reader variability and state-of-the-art methods.

Sign in / Sign up

Export Citation Format

Share Document