RescueNet: An unpaired GAN for brain tumor segmentation

Nema, Shubhangi; Dudhane, Akshay; Murala, Subrahmanyam; Naidu, Srivatsava

doi:10.1016/j.bspc.2019.101641

Cited by 136 publications

(54 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The comparison of these with the proposed method is given in Table 2 . The performance of deep learning based CNN [ 27 ] and RescueNet [ 32 ] is slightly better compared to the proposed method in terms of the evaluation metrics. But the proposed method has the advantage of less database requirement and improved running time.…”

Section: Resultsmentioning

confidence: 99%

“…The reconstruction is not the best due to the latent space loss and poor optimization. The training process in CNN and Generative Adversarial Network (GAN) [31][32][33][34] is both complex and time-consuming. So, in this paper, we suggest a simple autoencoder based training.…”

Section: Review Of Literaturementioning

confidence: 99%

“…In [ 31 , 32 ] the authors present the latest GAN model to segment the brain tumor. In [ 32 ] they developed a model named RescueNet which uses unpaired adversarial training to segment the whole tumor followed by core and enhance regions in a brain MRI scan. The authors claim the method can reduce the labeled data needed for deep learning.…”

Section: Review Of Literaturementioning

confidence: 99%

See 2 more Smart Citations

A dual autoencoder and singular value decomposition based feature optimization for the segmentation of brain tumor from MRI images

Aswani

Menaka

2021

BMC Med Imaging

View full text Add to dashboard Cite

Background The brain tumor is the growth of abnormal cells inside the brain. These cells can be grown into malignant or benign tumors. Segmentation of tumor from MRI images using image processing techniques started decades back. Image processing based brain tumor segmentation can be divided in to three categories conventional image processing methods, Machine Learning methods and Deep Learning methods. Conventional methods lacks the accuracy in segmentation due to complex spatial variation of tumor. Machine Learning methods stand as a good alternative to conventional methods. Methods like SVM, KNN, Fuzzy and a combination of either of these provide good accuracy with reasonable processing speed. The difficulty in processing the various feature extraction methods and maintain accuracy as per the medical standards still exist as a limitation for machine learning methods. In Deep Learning features are extracted automatically in various stages of the network and maintain accuracy as per the medical standards. Huge database requirement and high computational time is still poses a problem for deep learning. To overcome the limitations specified above we propose an unsupervised dual autoencoder with latent space optimization here. The model require only normal MRI images for its training thus reducing the huge tumor database requirement. With a set of normal class data, an autoencoder can reproduce the feature vector into an output layer. This trained autoencoder works well with normal data while it fails to reproduce an anomaly to the output layer. But a classical autoencoder suffer due to poor latent space optimization. The Latent space loss of classical autoencoder is reduced using an auxiliary encoder along with the feature optimization based on singular value decomposition (SVD). The patches used for training are not traditional square patches but we took both horizontal and vertical patches to keep both local and global appearance features on the training set. An Autoencoder is applied separately for learning both horizontal and vertical patches. While training a logistic sigmoid transfer function is used for both encoder and decoder parts. SGD optimizer is used for optimization with an initial learning rate of .001 and the maximum epochs used are 4000. The network is trained in MATLAB 2018a with a processor capacity of 3.7 GHz with NVIDIA GPU and 16 GB of RAM. Results The results are obtained using a patch size of 16 × 64, 64 × 16 for horizontal and vertical patches respectively. In Glioma images tumor is not grown from a point rather it spreads randomly. Region filling and connectivity operations are performed to get the final tumor segmentation. Overall the method segments Meningioma better than Gliomas. Three evaluation metrics are considered to measure the performance of the proposed system such as Dice Similarity Coefficient, Positive Predictive Value, and Sensitivity. Conclusion An unsupervised method for the segmentation of brain tumor from MRI images is proposed here. The proposed dual autoencoder with SVD based feature optimization reduce the latent space loss in the classical autoencoder. The proposed method have advantages in computational efficiency, no need of huge database requirement and better accuracy than machine learning methods. The method is compared Machine Learning methods Like SVM, KNN and supervised deep learning methods like CNN and commentable results are obtained.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Review Of Literaturementioning

confidence: 99%

Section: Review Of Literaturementioning

confidence: 99%

See 1 more Smart Citation

A dual autoencoder and singular value decomposition based feature optimization for the segmentation of brain tumor from MRI images

Aswani

Menaka

2021

BMC Med Imaging

View full text Add to dashboard Cite

show abstract

“…Earlier solutions to the challenge called brain tumor segmentation based on MRI data were summarized by Gordillo et al [4]. Recent solutions usually combine advanced (mostly unsupervised) image segmentation algorithms with semi-supervised supervised classification algorithms that cover the whole arsenal of machine learning techniques, namely: graph cut segmentation algorithm [5], superpixels combined with non-parametric classifiers [6], feature fusion combined with joint label fusion [7], texture feature and kernel sparse coding [8], Gaussian mixture models [9], fuzzy c-means clustering in semi-supervised context [10], fuzzy c-means clustering combined with region growing [11], AdaBoost classifier [12], extremely random trees [13] combined with superpixel level features [14], random forests [15,16] and ensemble of random forests [17], support vector machines [18], expert systems [19], convolutional neural network [20], deep neural networks [21], generative adversarial networks [22], and tumor growth model [23].…”

Section: Introductionmentioning

confidence: 99%

Brain Tumor Segmentation from Multi-spectral MR Image Data Using Random Forest Classifier

Csaholczi

Iclănzan

Kovács

et al. 2020

Neural Information Processing

View full text Add to dashboard Cite

The development of brain tumor segmentation techniques based on multi-spectral MR image data has relevant impact on the clinical practice via better diagnosis, radiotherapy planning and follow-up studies. This task is also very challenging due to the great variety of tumor appearances, the presence of several noise effects, and the differences in scanner sensitivity. This paper proposes an automatic procedure trained to distinguish gliomas from normal brain tissues in multi-spectral MRI data. The procedure is based on a random forest (RF) classifier, which uses 80 computed features beside the four observed ones, including morphological ones, gradients, and Gabor wavelet features. The intermediary segmentation outcome provided by the RF is fed to a twofold post-processing, which regularizes the shape of detected tumors and enhances the segmentation accuracy. The performance of the procedure was evaluated using the 274 records of the BraTS 2015 train data set. The achieved overall Dice scores between 85-86% represent highly accurate segmentation.

show abstract

“…Recently published solutions usually combine advanced unsupervised image segmentation algorithms with supervised and semi-supervised machine learning techniques. The wide spectrum of methodologies includes: active contour models combined with texture features [5], cellular automata combined with level sets [6], graph cut segmentation algorithm [7], superpixels combined with non-parametric classifiers [8], feature fusion combined with joint label fusion [9], texture feature and kernel sparse coding [10], Gaussian mixture models [11], [12], fuzzy cmeans clustering in semi-supervised context [13], [14], fuzzy c-means clustering combined with region growing [15], AdaBoost classifier [16], extremely random trees (ERT) [17] combined with superpixel level features [18], random forests [19], [20], [21] and ensemble of random forests [22], support vector machines [23], expert systems [24], convolutional neural network [25], [26], deep neural networks [27], [28], [29], [30], generative adversarial networks [31], and tumor growth model [32].…”

Section: Introductionmentioning

confidence: 99%

Brain Tumor Segmentation from Multi-Spectral Magnetic Resonance Image Data Using an Ensemble Learning Approach

Győrfi

Csaholczi

Fulop

et al. 2020

2020 IEEE International Conference on Systems, Man, and Cybernetics (SMC)

View full text Add to dashboard Cite

RescueNet: An unpaired GAN for brain tumor segmentation

Cited by 136 publications

References 15 publications

A dual autoencoder and singular value decomposition based feature optimization for the segmentation of brain tumor from MRI images

A dual autoencoder and singular value decomposition based feature optimization for the segmentation of brain tumor from MRI images

Brain Tumor Segmentation from Multi-spectral MR Image Data Using Random Forest Classifier

Brain Tumor Segmentation from Multi-Spectral Magnetic Resonance Image Data Using an Ensemble Learning Approach

Contact Info

Product

Resources

About