Glioma Tumors’ Classification Using Deep-Neural-Network-Based Features with SVM Classifier

Latif, Ghazanfar; Brahim, Ghassen Ben; Iskandar, Dedi; Bashar, Abul; Alghazo, Jaafar

doi:10.3390/diagnostics12041018

Cited by 63 publications

(27 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The authors demonstrated that their aggregation-based deep learning model performed much better than individual state-of-the-art CNNs, with an overall accuracy of 96.58%. The authors of [ 7 ] employed deep learning for the purpose of Glioma tumors. A multiclass model was built that used deep learning techniques for feature extraction and used SVMs for the purpose of classification.…”

Section: Introductionmentioning

confidence: 99%

Knee Osteoarthritis Detection and Severity Classification Using Residual Neural Networks on Preprocessed X-ray Images

et al. 2023

Self Cite

View full text Add to dashboard Cite

One of the most common and challenging medical conditions to deal with in old-aged people is the occurrence of knee osteoarthritis (KOA). Manual diagnosis of this disease involves observing X-ray images of the knee area and classifying it under five grades using the Kellgren–Lawrence (KL) system. This requires the physician’s expertise, suitable experience, and a lot of time, and even after that the diagnosis can be prone to errors. Therefore, researchers in the ML/DL domain have employed the capabilities of deep neural network (DNN) models to identify and classify KOA images in an automated, faster, and accurate manner. To this end, we propose the application of six pretrained DNN models, namely, VGG16, VGG19, ResNet101, MobileNetV2, InceptionResNetV2, and DenseNet121 for KOA diagnosis using images obtained from the Osteoarthritis Initiative (OAI) dataset. More specifically, we perform two types of classification, namely, a binary classification, which detects the presence or absence of KOA and secondly, classifying the severity of KOA in a three-class classification. For a comparative analysis, we experiment on three datasets (Dataset I, Dataset II, and Dataset III) with five, two, and three classes of KOA images, respectively. We achieved maximum classification accuracies of 69%, 83%, and 89%, respectively, with the ResNet101 DNN model. Our results show an improved performance from the existing work in the literature.

show abstract

Section: Introductionmentioning

confidence: 99%

Knee Osteoarthritis Detection and Severity Classification Using Residual Neural Networks on Preprocessed X-ray Images

et al. 2023

Self Cite

View full text Add to dashboard Cite

show abstract

“…Recently, the image processing field has been aided by Convolutional Neural Networks (CNN) [ 18 , 19 , 20 ]. An end-to-end system requiring minimal preprocessing results from the integration of the various image features and classifiers in CNN.…”

Section: Introductionmentioning

confidence: 99%

Diabetic Retinopathy Detection from Fundus Images of the Eye Using Hybrid Deep Learning Features

et al. 2022

Self Cite

View full text Add to dashboard Cite

Diabetic Retinopathy (DR) is a medical condition present in patients suffering from long-term diabetes. If a diagnosis is not carried out at an early stage, it can lead to vision impairment. High blood sugar in diabetic patients is the main source of DR. This affects the blood vessels within the retina. Manual detection of DR is a difficult task since it can affect the retina, causing structural changes such as Microaneurysms (MAs), Exudates (EXs), Hemorrhages (HMs), and extra blood vessel growth. In this work, a hybrid technique for the detection and classification of Diabetic Retinopathy in fundus images of the eye is proposed. Transfer learning (TL) is used on pre-trained Convolutional Neural Network (CNN) models to extract features that are combined to generate a hybrid feature vector. This feature vector is passed on to various classifiers for binary and multiclass classification of fundus images. System performance is measured using various metrics and results are compared with recent approaches for DR detection. The proposed method provides significant performance improvement in DR detection for fundus images. For binary classification, the proposed modified method achieved the highest accuracy of 97.8% and 89.29% for multiclass classification.

show abstract

“…In the second stage, proposed techniques for the segmentation of the tumorous image were explained. In the third stage, proposed significant contributions for the classification of tumorous images into the four Glioma classes; Necrosis, Edema, Enhancing, and Non-enhancing were detailed [ 5 ].…”

Section: Introductionmentioning

confidence: 99%

DeepTumor: Framework for Brain MR Image Classification, Segmentation and Tumor Detection

Latif

2022

Diagnostics

Self Cite

View full text Add to dashboard Cite

The proper segmentation of the brain tumor from the image is important for both patients and medical personnel due to the sensitivity of the human brain. Operation intervention would require doctors to be extremely cautious and precise to target the brain’s required portion. Furthermore, the segmentation process is also important for multi-class tumor classification. This work primarily concentrated on making a contribution in three main areas of brain MR Image processing for classification and segmentation which are: Brain MR image classification, tumor region segmentation and tumor classification. A framework named DeepTumor is presented for the multistage-multiclass Glioma Tumor classification into four classes; Edema, Necrosis, Enhancing and Non-enhancing. For the brain MR image binary classification (Tumorous and Non-tumorous), two deep Convolutional Neural Network) CNN models were proposed for brain MR image classification; 9-layer model with a total of 217,954 trainable parameters and an improved 10-layer model with a total of 80,243 trainable parameters. In the second stage, an enhanced Fuzzy C-means (FCM) based technique is proposed for the tumor segmentation in brain MR images. In the final stage, an enhanced CNN model 3 with 11 hidden layers and a total of 241,624 trainable parameters was proposed for the classification of the segmented tumor region into four Glioma Tumor classes. The experiments are performed using the BraTS MRI dataset. The experimental results of the proposed CNN models for binary classification and multiclass tumor classification are compared with the existing CNN models such as LeNet, AlexNet and GoogleNet as well as with the latest literature.

show abstract

Glioma Tumors’ Classification Using Deep-Neural-Network-Based Features with SVM Classifier

Cited by 63 publications

References 25 publications

Knee Osteoarthritis Detection and Severity Classification Using Residual Neural Networks on Preprocessed X-ray Images

Knee Osteoarthritis Detection and Severity Classification Using Residual Neural Networks on Preprocessed X-ray Images

Diabetic Retinopathy Detection from Fundus Images of the Eye Using Hybrid Deep Learning Features

DeepTumor: Framework for Brain MR Image Classification, Segmentation and Tumor Detection

Contact Info

Product

Resources

About