Delve into Multiple Sclerosis (MS) lesion exploration: A modified attention U-Net for MS lesion segmentation in Brain MRI

Hashemi, Maryam; Akhbari, Mahsa; Jutten, Christian

doi:10.1016/j.compbiomed.2022.105402

Cited by 33 publications

(21 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This section provides the segmentation efficiency of the proposed NICNN over the existing UNET-based model [12,14,27]. The proposed model is implemented using Python 3 and MATLAB framework using Windows 10 operating system running on I-7 quad-core processor, with 16GB RAM and CUDA-enabled 4GB GPU.…”

Section: Resultsmentioning

confidence: 99%

“…When applying the transformation, it exhibits a multi-resolution pattern; thus, enhancing segmentation outcomes with varying lesion sizes. In [14], designed an MS lesion segmentation framework using T2 and FLAIR MRI scans. They modified the attention UNET (MAUNET) and also modified the UNET (MUNET) by introducing additional preprocessing and loss functions [15].…”

Section: 1mentioning

confidence: 99%

“…This section provides the architecture configuration of NICNN as shown in Table 1 and graphical representation of segmentation outcome of proposed NICNN over existing segmentation model namely MUNET [14] and MAUNET [14]. The layer size, parameter size, training and testing time is shown in Table 1.…”

Section: Segmentation Outcomementioning

confidence: 99%

See 2 more Smart Citations

Noise Invariant Convolution Neural Network for Segmentation of Multiple Sclerosis Lesions from Brain Magnetic Resonance Imaging

Swetha¹,

Aditya²

2022

Int. J. Onl. Eng.

View full text Add to dashboard Cite

The objective of the research work is to accurately segment multiple sclerosis (MS) lesions in brain Magnetic Resonance Imaging (MRI) of varying sizes and also to classify its types. Designing effective automatic segmentation and classification tool aid the doctors in better understanding MS lesion progressions. In meeting research challenges, this paper presents Noise Invariant Convolution Neural Network (NICNN). The NICNN model is efficient in the detection and segmentation of MS lesions of varying sizes in comparison with standard CNN-based segmentation methods. Further, this paper introduced a new cross-validation scheme to address the class imbalance issue by selecting effective features for classifying the type of MS lesion. The experiment outcome shows the proposed method provides improved Dice Similarity Coefficient (DSC), Positive Predicted Value (PPV), and True Positive Rate (TPR) value compared to the state-of-art CNN-based MS lesion segmentation method. Further, achieves better accuracy in classifying MS lesion types compared to standard MS lesion type classification models.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: 1mentioning

confidence: 99%

See 1 more Smart Citation

Noise Invariant Convolution Neural Network for Segmentation of Multiple Sclerosis Lesions from Brain Magnetic Resonance Imaging

Swetha¹,

Aditya²

2022

Int. J. Onl. Eng.

View full text Add to dashboard Cite

show abstract

“…Early stopping (patience=50) was exploited to prevent overfitting as well. Hashemi et al ( 2022 ) used the sum of dice loss with a 1.5 coefficient and binary cross entropy loss as a custom loss function for MS lesion segmentation. Similarly, in this study, a hybrid loss function consisting of binary focal loss and dice loss [dice loss + (1 × binary focal loss)] was employed in order to handle unbalanced labeled data between lesion and background since lesion pixels constitute a minor portion of the whole image.…”

Section: Methodsmentioning

confidence: 99%

New MS lesion segmentation with deep residual attention gate U-Net utilizing 2D slices of 3D MR images

Sarica

Şeker²

2022

Front. Neurosci.

View full text Add to dashboard Cite

Multiple sclerosis (MS) is an autoimmune disease that causes lesions in the central nervous system of humans due to demyelinating axons. Magnetic resonance imaging (MRI) is widely used for monitoring and measuring MS lesions. Automated methods for MS lesion segmentation have usually been performed on individual MRI scans. Recently, tracking lesion activity for quantifying and monitoring MS disease progression, especially detecting new lesions, has become an important biomarker. In this study, a unique pipeline with a deep neural network that combines U-Net, attention gate, and residual learning is proposed to perform better new MS lesion segmentation using baseline and follow-up 3D FLAIR MR images. The proposed network has a similar architecture to U-Net and is formed from residual units which facilitate the training of deep networks. Networks with fewer parameters are designed with better performance through the skip connections of U-Net and residual units, which facilitate information propagation without degradation. Attention gates also learn to focus on salient features of the target structures of various sizes and shapes. The MSSEG-2 dataset was used for training and testing the proposed pipeline, and the results were compared with those of other proposed pipelines of the challenge and experts who participated in the same challenge. According to the results over the testing set, the lesion-wise F1 and dice scores were obtained as a mean of 48 and 44.30%. For the no-lesion cases, the number of tested and volume of tested lesions were obtained as a mean of 0.148 and 1.488, respectively. The proposed pipeline outperformed 22 proposed pipelines and ranked 8th in the challenge.

show abstract

“…

where TP denotes the true positive voxels, TN denotes the true negative voxels, FP denotes the false positive voxels, and FN denotes the false negative voxels. We use the 4-union (IOU), DSC value, and Hausdorff distance [ 58 ]. Here, we defined IOU as

…”

Section: Methodsmentioning

confidence: 99%

SegChaNet: A Novel Model for Lung Cancer Segmentation in CT Scans

Çifçi

2022

Applied Bionics and Biomechanics

View full text Add to dashboard Cite

Accurate lung tumor identification is crucial for radiation treatment planning. Due to the low contrast of the lung tumor in computed tomography (CT) images, segmentation of the tumor in CT images is challenging. This paper effectively integrates the U-Net with the channel attention module (CAM) to segment the malignant lung area from the surrounding chest region. The SegChaNet method encodes CT slices of the input lung into feature maps utilizing the trail of encoders. Finally, we explicitly developed a multiscale, dense-feature extraction module to extract multiscale features from the collection of encoded feature maps. We have identified the segmentation map of the lungs by employing the decoders and compared SegChaNet with the state-of-the-art. The model has learned the dense-feature extraction in lung abnormalities, while iterative downsampling followed by iterative upsampling causes the network to remain invariant to the size of the dense abnormality. Experimental results show that the proposed method is accurate and efficient and directly provides explicit lung regions in complex circumstances without postprocessing.

show abstract

Delve into Multiple Sclerosis (MS) lesion exploration: A modified attention U-Net for MS lesion segmentation in Brain MRI

Cited by 33 publications

References 17 publications

Noise Invariant Convolution Neural Network for Segmentation of Multiple Sclerosis Lesions from Brain Magnetic Resonance Imaging

Noise Invariant Convolution Neural Network for Segmentation of Multiple Sclerosis Lesions from Brain Magnetic Resonance Imaging

New MS lesion segmentation with deep residual attention gate U-Net utilizing 2D slices of 3D MR images

SegChaNet: A Novel Model for Lung Cancer Segmentation in CT Scans

Contact Info

Product

Resources

About