MDRN: Multi-distillation residual network for efficient MR image super-resolution
Super-resolution (SR) of magnetic resonance imaging (MRI) is gaining increasing attention for being able to provide detailed anatomical information. However, current SR methods often use the complex convolutional network for feature extraction, which is difficult to train and not suitable for limite...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
AIMS Press
2024-10-01
|
| Series: | Mathematical Biosciences and Engineering |
| Subjects: | |
| Online Access: | https://www.aimspress.com/article/doi/10.3934/mbe.2024326 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832590852541644800 |
|---|---|
| author | Liwei Deng Jingyi Chen Xin Yang Sijuan Huang |
| author_facet | Liwei Deng Jingyi Chen Xin Yang Sijuan Huang |
| author_sort | Liwei Deng |
| collection | DOAJ |
| description | Super-resolution (SR) of magnetic resonance imaging (MRI) is gaining increasing attention for being able to provide detailed anatomical information. However, current SR methods often use the complex convolutional network for feature extraction, which is difficult to train and not suitable for limited computation resources in the medical scenario. To tackle these bottlenecks, we propose a multi-distillation residual network (MDRN) for more differential feature refinement, which has a superior trade-off between reconstruction accuracy and computation cost. Specifically, a novel feature multi-distillation residual block with a contrast-aware channel attention module was designed to make the residual features more focused on low-vision information, which maximizes the power of MDRN. Comprehensive experiments demonstrate the superiority of our MDRN over state-of-the-art methods in reconstruction quality and efficiency. Our method outperforms other existing methods in peak signal-noise ratio by up to 0.44–1.82 dB in 4× scale when GPU memory and runtime are lower than in other SR methods. The source code will be available at https://github.com/Jennieyy/MDRN. |
| format | Article |
| id | doaj-art-e718bdfa75414ea1b2d5e0e444d561fc |
| institution | Kabale University |
| issn | 1551-0018 |
| language | English |
| publishDate | 2024-10-01 |
| publisher | AIMS Press |
| record_format | Article |
| series | Mathematical Biosciences and Engineering |
| spelling | doaj-art-e718bdfa75414ea1b2d5e0e444d561fc2025-01-23T07:48:00ZengAIMS PressMathematical Biosciences and Engineering1551-00182024-10-0121107421743410.3934/mbe.2024326MDRN: Multi-distillation residual network for efficient MR image super-resolutionLiwei Deng0Jingyi Chen1Xin Yang2Sijuan Huang3Heilongjiang Provincial Key Laboratory of Complex Intelligent System and Integration, School of Automation, Harbin University of Science and Technology, Harbin 150080, Heilongjiang, ChinaHeilongjiang Provincial Key Laboratory of Complex Intelligent System and Integration, School of Automation, Harbin University of Science and Technology, Harbin 150080, Heilongjiang, ChinaGuangdong Esophageal Cancer Institute, Guangzhou 510060, Guangdong, ChinaState Key Laboratory of Oncology in South China, Guangdong Key Laboratory of Nasopharyngeal Carcinoma Diagnosis and Therapy, Guangdong Provincial Clinical Research Center for Cancer, Sun Yat-sen University Cancer Center, Guangzhou 510060, Guangdong, ChinaSuper-resolution (SR) of magnetic resonance imaging (MRI) is gaining increasing attention for being able to provide detailed anatomical information. However, current SR methods often use the complex convolutional network for feature extraction, which is difficult to train and not suitable for limited computation resources in the medical scenario. To tackle these bottlenecks, we propose a multi-distillation residual network (MDRN) for more differential feature refinement, which has a superior trade-off between reconstruction accuracy and computation cost. Specifically, a novel feature multi-distillation residual block with a contrast-aware channel attention module was designed to make the residual features more focused on low-vision information, which maximizes the power of MDRN. Comprehensive experiments demonstrate the superiority of our MDRN over state-of-the-art methods in reconstruction quality and efficiency. Our method outperforms other existing methods in peak signal-noise ratio by up to 0.44–1.82 dB in 4× scale when GPU memory and runtime are lower than in other SR methods. The source code will be available at https://github.com/Jennieyy/MDRN.https://www.aimspress.com/article/doi/10.3934/mbe.2024326super-resolutionmri reconstructionfeature distillationmedical image processing |
| spellingShingle | Liwei Deng Jingyi Chen Xin Yang Sijuan Huang MDRN: Multi-distillation residual network for efficient MR image super-resolution Mathematical Biosciences and Engineering super-resolution mri reconstruction feature distillation medical image processing |
| title | MDRN: Multi-distillation residual network for efficient MR image super-resolution |
| title_full | MDRN: Multi-distillation residual network for efficient MR image super-resolution |
| title_fullStr | MDRN: Multi-distillation residual network for efficient MR image super-resolution |
| title_full_unstemmed | MDRN: Multi-distillation residual network for efficient MR image super-resolution |
| title_short | MDRN: Multi-distillation residual network for efficient MR image super-resolution |
| title_sort | mdrn multi distillation residual network for efficient mr image super resolution |
| topic | super-resolution mri reconstruction feature distillation medical image processing |
| url | https://www.aimspress.com/article/doi/10.3934/mbe.2024326 |
| work_keys_str_mv | AT liweideng mdrnmultidistillationresidualnetworkforefficientmrimagesuperresolution AT jingyichen mdrnmultidistillationresidualnetworkforefficientmrimagesuperresolution AT xinyang mdrnmultidistillationresidualnetworkforefficientmrimagesuperresolution AT sijuanhuang mdrnmultidistillationresidualnetworkforefficientmrimagesuperresolution |