3D Multi-Scale Convolutional Networks for Glioma Grading Using MR Images

Ge, Chenjie; Qu, Qixun; Gu, Irene Yu‐Hua; Jakola, Asgeir Store

doi:10.1109/icip.2018.8451682

Cited by 30 publications

(22 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Observing Tables 4 and 5, it is shown that the proposed method is better than those in [19,30] in terms of test accuracy. It is also indicated that the proposed method has reached high performance as comparing with the methods in [13,14,17,18], noting the results were obtained from different datasets with different number of patients.…”

Section: Impact Of Adding Gan-augmented Images In the Trainingmentioning

confidence: 91%

“…All experiments were done on a workstation with Intel-i7 3.40GHz CPU, 48G RAM and an NVIDIA Titan Xp 12GB GPU. Hyperparameter settings were as follows: for TCGA dataset, pretraining was applied to GAN augmented images using 100 epochs with the learning rate 1e-4 for epochs∈ [1,30], 1e-5 for epochs∈ [31,60], and 1e-6 for epochs∈[61,100]. Refined training was then applied to the original images using 50 epochs with the learning rate 1e-5.…”

Section: Setup Datasets and Metricsmentioning

confidence: 99%

“…It is worth noting that results from [13,14,17,18] in Table 4, and the results from [29] in Table 5 can only be used as an indication since they were applied in different datasets. [30] used a very different approach as our proposed scheme here, which applied a 3D multiscale CNN network directly on 3D T1ce MR images, with fully annotated training dataset.…”

Section: Impact Of Adding Gan-augmented Images In the Trainingmentioning

confidence: 99%

See 2 more Smart Citations

Deep semi-supervised learning for brain tumor classification

Jakola

et al. 2020

BMC Med Imaging

Self Cite

View full text Add to dashboard Cite

Background: This paper addresses issues of brain tumor, glioma, classification from four modalities of Magnetic Resonance Image (MRI) scans (i.e., T1 weighted MRI, T1 weighted MRI with contrast-enhanced, T2 weighted MRI and FLAIR). Currently, many available glioma datasets often contain some unlabeled brain scans, and many datasets are moderate in size. Methods: We propose to exploit deep semi-supervised learning to make full use of the unlabeled data. Deep CNN features were incorporated into a new graph-based semi-supervised learning framework for learning the labels of the unlabeled data, where a new 3D-2D consistent constraint is added to make consistent classifications for the 2D slices from the same 3D brain scan. A deep-learning classifier is then trained to classify different glioma types using both labeled and unlabeled data with estimated labels. To alleviate the overfitting caused by moderate-size datasets, synthetic MRIs generated by Generative Adversarial Networks (GANs) are added in the training of CNNs. Results: The proposed scheme has been tested on two glioma datasets, TCGA dataset for IDH-mutation prediction (molecular-based glioma subtype classification) and MICCAI dataset for glioma grading. Our results have shown good performance (with test accuracies 86.53% on TCGA dataset and 90.70% on MICCAI dataset). Conclusions: The proposed scheme is effective for glioma IDH-mutation prediction and glioma grading, and its performance is comparable to the state-of-the-art.

show abstract

Section: Impact Of Adding Gan-augmented Images In the Trainingmentioning

confidence: 91%

Section: Setup Datasets and Metricsmentioning

confidence: 99%

Section: Impact Of Adding Gan-augmented Images In the Trainingmentioning

confidence: 99%

See 1 more Smart Citation

Deep semi-supervised learning for brain tumor classification

Jakola

et al. 2020

BMC Med Imaging

Self Cite

View full text Add to dashboard Cite

show abstract

“…DNNs make it possible to solve several practical tasks in real life, such as detecting cancer [17,31,32,84] or Alzheimer's [9,60,61]. Furthermore, DNN studies of facial properties can reveal several diseases, sexual orientation, IQ, and political preferences.…”

Section: Relations Between Ai and Ami Super-ai And Super-amimentioning

confidence: 99%

“…In [17] a residual CNN was proposed in combination with several datasets along with data augmentation that yielded an IDH prediction accuracy of 87.6% on the test set. In [32] a 3D multiscale CNN and fusion were used to exploit multiscale features on MRIs with saliency-enhanced tumour regions that resulted in an 89.47% test accuracy for low/high-grade gliomas. Finally, [84] proposed using an SVM classifier on multimodal MRIs for predicting gliomas with extensive subcategories, like IDH wild, IDH mutant with TP53 wild/mutate subcategories and more with relatively good results.…”

Section: Ami For Assisting Medical Diagnosis: Brain Tumours and Alzhementioning

confidence: 99%

Artificial intelligence and ambient intelligence

Gams

Härmä

et al. 2019

AIS

Self Cite

View full text Add to dashboard Cite

Ambient intelligence (AmI) is intrinsically and thoroughly connected with artificial intelligence (AI). Some even say that it is, in essence, AI in the environment. AI, on the other hand, owes its success to the phenomenal development of the information and communication technologies (ICTs), based on principles such as Moore's law. In this paper we give an overview of the progress in AI and AmI interconnected with ICT through information-society laws, superintelligence, and several related disciplines, such as multi-agent systems and the Semantic Web, ambient assisted living and e-healthcare, AmI for assisting medical diagnosis, ambient intelligence for e-learning and ambient intelligence for smart cities. Besides a short history and a description of the current state, the frontiers and the future of AmI and AI are also considered in the paper.

show abstract

Applying software‐defined ambient intelligence techniques for cognitive and emotional awareness in the Internet of Things

Azeez Syed,

Nag,

Basu

et al. 2023

Int J Communication

View full text Add to dashboard Cite

SummarySmart “discrete” technologies are taking over everyday activities with an intended shift towards complete automation. With this, ambient intelligence (AmI) has come under the limelight, to make these routines subtle and automated, enabled through the widespread adoption of the ubiquitous and pervasive Internet of Things (IoT). AmI is context‐driven depending largely on user input through various interfaces, which is nothing but the interplay between IoT and artificial intelligence (AI). Despite extensive incorporation of personal digital assistants and smart home systems, AI is not yet capable of accurately interpreting and responding to human emotions. This paper aims to present one such application to gauge the emotional and cognitive state of the user and respond appropriately, be it ambient lighting or an emergency SOS. The proposed model extracts network packets to work with audio messages on the real‐time transport protocol (RTP) layer and presents a lightweight multilayer perceptron (MLP)‐based speech emotion recognition (SER) model for emotional analysis with an accuracy of 80.52%. It also features a simple confidence rating system for further optimization.

show abstract

3D Multi-Scale Convolutional Networks for Glioma Grading Using MR Images

Cited by 30 publications

References 6 publications

Deep semi-supervised learning for brain tumor classification

Deep semi-supervised learning for brain tumor classification

Artificial intelligence and ambient intelligence

Applying software‐defined ambient intelligence techniques for cognitive and emotional awareness in the Internet of Things

Contact Info

Product

Resources

About