Book Chaptershttp://hdl.handle.net/20.500.14131/162024-03-28T14:59:17Z2024-03-28T14:59:17ZBrain stroke detection from computed tomography images using deep learning algorithmsAykut DikerAbdullah ElenSubasi, Abdulhamithttp://hdl.handle.net/20.500.14131/14972024-03-12T10:57:07Z2023-01-01T00:00:00ZBrain stroke detection from computed tomography images using deep learning algorithms
Aykut Diker; Abdullah Elen; Subasi, Abdulhamit
Stroke is one of the common causes of death worldwide. Stroke is the inability of a focus to be fed in the brain due to clogged or bleeding of the vessels feeding the brain. Because early stroke treatment and diagnosis are related to a favorable patient outcome, time is a critical aspect of successful stroke treatment. In this chapter, we examine the stroke classification from Brain Stroke CT Dataset, with deep learning architectures. In the experimental study, a total of 2501 brain stroke computed tomography (CT) images were used for testing and training. For this purpose, numerus widely known pretrained convolutional neural networks (CNNs) such as GoogleNet, AlexNet, VGG-16, VGG-19, and Residual CNN were used to classify brain stroke CT images as normal and as stroke. Several performance metrics such as accuracy (ACC), specificity (SPE), sensitivity (SEN), and F-score are used to evaluate the performances of the classifier. The best classification results are achieved by VGG-19 with ACC 97.06%, SEN 97.41%, SPE 96.49%, and F-score 96.95%.
2023-01-01T00:00:00ZBreast tumor detection in ultrasound images using artificial intelligenceOmkar ModiSubasi, Abdulhamithttp://hdl.handle.net/20.500.14131/14962024-03-12T10:56:07Z2023-01-01T00:00:00ZBreast tumor detection in ultrasound images using artificial intelligence
Omkar Modi; Subasi, Abdulhamit
Leveraging artificial intelligence (AI) for categorizing breast tumors as malignant or benign from breast ultrasound images can provide an effective and relatively low-cost method for the diagnosis of breast cancer. Presently, many machine learning (ML) and deep learning (DL) algorithms have been used for early-stage breast cancer detection. AI algorithms have shown promising results in breast cancer detection tasks. The use of deep convolutional neural network approaches has provided solutions for the efficient analysis of breast ultrasound images. Convolutional neural network (CNN) models analyze the image data in multiple layers and extract features which helps in better feature extractions and better performance in comparison to the conventional ML algorithms. Apart from conventional learning algorithms, we use the transfer learning technique which uses knowledge from its previous training in another related problem set. In this chapter, we have demonstrated the use of DL models through transfer learning, deep feature extraction, machine learning models, and comparison of their performances.
2023-01-01T00:00:00ZBreast cancer detection from mammograms using artificial intelligenceSubasi, AbdulhamitAayush Dinesh KandpalKolla Anant RajUlas Bagcihttp://hdl.handle.net/20.500.14131/14952024-03-12T10:50:16Z2023-01-01T00:00:00ZBreast cancer detection from mammograms using artificial intelligence
Subasi, Abdulhamit; Aayush Dinesh Kandpal; Kolla Anant Raj; Ulas Bagci
Breast cancer is one of the fastest-growing forms of cancer in the world today. Breast cancer is primarily found in women, and its frequency has been gaining significantly in the last few years. The key to tackle the rising cases of breast cancer is early detection. Many studies have shown that early detection significantly reduces the mortality rate of those affected. Machine learning and deep learning techniques have been adopted in the present scenario to help detect breast cancer in an early stage. Deep learning models such as the convolutional neural networks (CNNs) are suited explicitly to image data and overcome the drawbacks of machine learning models. To improve upon conventional approaches, we apply deep CNNs for automatic feature extraction and classifier building. In this chapter, we have demonstrated thoroughly the use of deep learning models through transfer learning, deep feature extraction, and machine learning models. Computer-aided detection or diagnosis systems have recently been developed to help health-care professionals increase diagnosis accuracy. This chapter presents early breast cancer detection from mammograms using artificial intelligence (AI). Various models have been presented along with an in-depth comparative analysis of the different state-of-the-art architectures, custom CNN networks, and classifiers trained on features extracted from pretrained networks. Our findings have indicated that deep learning models can achieve training accuracies of up to 99%, while both validation and test accuracies up to 96%. We conclude by suggesting various improvements that could be made to existing architectures and how AI techniques could help further improve and help in the early detection of breast cancer.
2023-01-01T00:00:00ZArtificial intelligence-based retinal disease classification using optical coherence tomography imagesSohan PatnaikSubasi, Abdulhamithttp://hdl.handle.net/20.500.14131/14942024-03-12T10:45:01Z2023-01-01T00:00:00ZArtificial intelligence-based retinal disease classification using optical coherence tomography images
Sohan Patnaik; Subasi, Abdulhamit
Optical coherence tomography (OCT) is a noninvasive imaging technology used to obtain high-resolution cross-sectional images of the retina. The layers within the retina can be differentiated and retinal thickness can be measured to aid in the early detection and diagnosis of retinal diseases and conditions. Notwithstanding the proven utility of OCT images, diagnosing large datasets of OCT images using the manual method still remains a challenge. In this chapter, we propose a deep learning-based approach, namely, the use of convolutional neural networks (CNN) and some pretrained image classification models on top of CNNs to get a proper and faster diagnosis of the OCT images. We also experiment with the features extracted using pretrained image classification models. Mainly three diseases—drusen, diabetic macular edema, choroidal neovascularization are addressed in this study. Our technique achieves an accuracy score of 0.9948 and an F1 score of 0.9948 on the test set. The outcomes suggest that the proposed model could serve as an analytic module to alert users and/or medical experts when retinal diseases are suspected.
2023-01-01T00:00:00Z