scholarly journals Compressive Sensing Magnetic Resonance Image Reconstruction and Denoising using Convolutional Neural Network

2022 ◽  
Vol 2161 (1) ◽  
pp. 012036
Author(s):  
Ram Singh ◽  
Lakhwinder Kaur
Keyword(s):  
Magnetic Resonance ◽  
Image Reconstruction ◽  
Image Quality ◽  
Data Acquisition ◽  
Compressive Sensing ◽  
Magnetic Resonance Image ◽  
Motion Artifacts ◽  
Acquisition Time ◽  
High Quality ◽  
Mr Image

Abstract Restoration of high-quality brain Magnetic Resonance Image (MRI) from the sparse under-sampled complex k-space signal is a widely studied ill-posed inverse transform problem. A deep learning-based data-adaptive and data-driven convolutional technique has been proposed for high-quality MRI recovery from its under-sampled complex domain k-space signal. The uniform subsampling process is very slow in phase-encoding to generate high-resolution images. The longer scan times degrade the perceptual image quality. Various factors contribute to image degradation during data acquisition such as the inception of body motion artifacts, the thermal energy effects of the body, and random noise artifacts due to voltage fluctuations. Keeping in view the patient’s critical condition and comfort, longer scan times are not preferred in practice. To reduce the image acquisition time, noise levels, and motion artifacts in the MR images, Compressive Sensing (CS) provides an accelerated way to reconstructs the high-quality MR image from very limited signal measurements acquired much below the Nyquist rate. However, such data acquisition strategies require advanced computer algorithms for the reconstruction of high-quality MRI from the undersampled MRI data. An improved CNN-based MRI reconstructed algorithm has been presented in this paper which shows better performance to reconstruct high-quality MRI than similar other MR image reconstruction algorithms. The performance of the proposed algorithm is measured by image quality checking tools such as normalized-MSE, PSNR, and SSIM.

scholarly journals Noise-residue learning convolutional network model for magnetic resonance image enhancement

2021 ◽  
Vol 2089 (1) ◽  
pp. 012029
Author(s):  
Ram Singh ◽  
Lakhwinder Kaur
Keyword(s):  
Magnetic Resonance ◽  
High Resolution ◽  
Magnetic Resonance Image ◽  
Motion Artifacts ◽  
Tumor Segmentation ◽  
X Rays ◽  
High Quality ◽  
High Resolution Mri ◽  
Segmentation Image ◽  
Filter Noise

Abstract Magnetic Resonance Image (MRI) is an important medical image acquisition technique used to acquire high contrast images of human body anatomical structures and soft tissue organs. MRI system does not use any harmful radioactive ionized material like x-rays and computerized tomography (CT) imaging techniques. High-resolution MRI is desirable in many clinical applications such as tumor segmentation, image registration, edges & boundary detection, and image classification. During MRI acquisition, many practical constraints limit the MRI quality by introducing random Gaussian noise and some other artifacts by the thermal energy of the patient body, random scanner voltage fluctuations, body motion artifacts, electronics circuits impulse noise, etc. High-resolution MRI can be acquired by increasing scan time, but considering patient comfort, it is not preferred in practice. Hence, postacquisition image processing techniques are used to filter noise contents and enhance the MRI quality to make it fit for further image analysis tasks. The main motive of MRI enhancement is to reconstruct a high-quality MRI while improving and retaining its important features. The new deep learning image denoising and artifacts removal methods have shown tremendous potential for high-quality image reconstruction from noise degraded MRI while preserving useful image information. This paper presents a noise-residue learning convolution neural network (CNN) model to denoise and enhance the quality of noise-corrupted low-resolution MR images. The proposed technique shows better performance in comparison with other conventional MRI enhancement methods. The reconstructed image quality is evaluated by the peak-signal-to-noise ratio (PSNR) and structural similarity index (SSIM) metrics by optimizing information loss in reconstructed MRI measured in mean squared error (MSE) metric.

scholarly journals To Analyse the Effect of Relaxation Type on Magnetic Resonance Image Compression Using Compressive Sensing

2021 ◽  
Vol 17 (04) ◽  
pp. 21
Author(s):  
Vivek Upadhyaya ◽  
Mohammad Salim
Keyword(s):  
Magnetic Resonance ◽  
Compressive Sensing ◽  
Magnetic Resonance Image ◽  
Similarity Index ◽  
Structural Similarity ◽  
Scanning Speed ◽  
Proton Density ◽  
Body Parts ◽  
Mr Image ◽  
Reconstruction Process

<span>Medical Imaging and scanning technologies are used to provide better resolution of body and tissues. To achieve a better quality Magnetic Resonance (MR) image with a minimum duration of processing time is a tedious task. So our purpose in this paper is to find out a solution that can minimize the reconstruction time of an MRI signal. </span><span>Compressive sensing can be used to accelerate Magnetic Resonance Image (MRI) acquisition by acquiring fewer data through the under-sampling of k-space, so it can be used to minimize the time. But according to the relaxation time, we can further classify the MRI signal into T1, T2, and Proton Density (PD) weighted images. These weighted images represent different signal intensities for different types of tissues and body parts. It also affects the reconstruction process conducted by using the Compressive Sensing Approach. This study is based on finding out the effect of T1, T2, and Proton Density (PD) weighted images on the reconstruction process as well as various image quality parameters like MSE, PSNR, &amp; SSIM also calculated to analyze this effect. Meanwhile, we can analyze how many samples are enough to reconstruct the MR image so the problem associated with time and scanning speed can be reduced up to an extent. In this paper, we got the Structural Similarity Index Measure (SSIM) value up to 0.89 &amp; PSNR value 37.83451 dB at an 85 % compression ratio for the T2 weighted image. </span>

Technical Aspects on Magnetic Resonance Imaging of the Spine at 1.5 Tesla

1987 ◽  
Vol 28 (4) ◽  
pp. 375-381 ◽  
Author(s):  
S. L. Holtås ◽  
D. B. Plewes ◽  
J. H. Simon ◽  
S. Ekholm ◽  
D. K. Kido ◽  
...  
Keyword(s):  
Magnetic Resonance Imaging ◽  
Magnetic Resonance ◽  
Image Quality ◽  
Surface Coil ◽  
Motion Artifacts ◽  
The Body ◽  
Acquisition Time ◽  
Slice Thickness ◽  
Resonance Imaging ◽  
Technical Aspects

Technical aspects on surface coil magnetic resonance imaging of the spine using a superconducting system with a field strength of 1.5 tesla are described. By using a flat surface coil instead of the body coil the image quality was markedly improved and the signal-to-noise ratio (S/N) was increased approximately 2.6 times. Small voxels resulted in low S/N. The best image quality was achieved with a slice thickness of 5 mm, a field of view of 20 to 24 cm and a matrix of 256×256. Interleaved slices provided superior image quality compared with contiguous slices at the expense of acquisition time. For sagittal images the phase encoding gradient should be in the cranio-caudal direction to minimize motion artifacts. To obtain T1 and T2 images of high quality, spin echo pulse sequences with TR 600/TE 20 ms and TR 2000/TE 40 to 80 ms are useful.

Deep learning in magnetic resonance image reconstruction

2021 ◽  
Author(s):  
Shekhar S Chandra ◽  
Marlon Bran Lorenzana ◽  
Xinwen Liu ◽  
Siyu Liu ◽  
Steffen Bollmann ◽  
...  
Keyword(s):  
Deep Learning ◽  
Magnetic Resonance ◽  
Image Reconstruction ◽  
Magnetic Resonance Image

An Efficient Light-Weight Network for Fast Reconstruction on MR Images

2021 ◽  
Vol 17 ◽  
Author(s):  
Bowen Zhen ◽  
Yingjie Zheng ◽  
Bensheng Qiu
Keyword(s):  
Feature Extraction ◽  
Image Reconstruction ◽  
Data Consistency ◽  
Training Procedure ◽  
Light Weight ◽  
Mr Images ◽  
High Quality ◽  
Mr Image ◽  
Generalization Ability ◽  
Mr Image Reconstruction

Background: In recent years, deep learning (DL) algorithms have emerged in endlessly and achieved impressive performance, which makes it possible to accelerate magnetic resonance (MR) image reconstruction with DL instead of compressed sensing (CS) methods. However, a DL-based MR image reconstruction method has always suffered from its heavy learning parameters and poor generalization ability so far. Therefore, an efficient light-weight network is still in desperate need of fast MR image reconstruction. Methods: We propose an efficient and light-weight MR reconstruction network (named RecNet) that uses a Convolutional Neural Network (CNN) to fast reconstruct high-quality MR images. Specifically, the network is composed of cascade modules, and each cascade module is further divided into feature extraction blocks and a data consistency layer. The feature extraction block can not only effectively extract the features of MR images, but also do not introduce too many parameters for the whole network. To stabilize the training procedure, the correction information of image frequency is adopted in the data consistency (DC) layer. Results: We have evaluated RecNet on a public dataset and the results show that the image quality reconstructed by RecNet is the best on the peak a signal-to-noise ratio (PSNR) and structural similarity index (SSIM) evaluation standards. In addition, the pre-trained RecNet can also reconstruct high-quality MR images on an unseen dataset. Conclusion: The results demonstrate that the RecNet has superior reconstruction ability in various metrics than comparative methods. The RecNet can quickly generate high-quality MR images in fewer parameters. Furthermore, the RecNet has an excellent generalization ability on pathological images and different sampling rates data.

Sign in / Sign up

Export Citation Format

Share Document