2023 · 44 citations · 10 references
ComputingComputer ScienceOncologyDiagnostic ImagingMachine LearningNeuro-oncologyBrain TumorsDeep LearningMedical Image AnalysisMedical ImagingNeural NetworksTumor DetectionAccuracyComparative Data AnalysisConvolutional Neural Network DetectionDeep Learning In RadiologyDeep Learning-based DiagnosisDeep Learning-based Lesion Classification
This research examines using MRI images for brain tumour identification in deep learning-based tumour prediction models. Four well-known deep learning architectures’ accuracy and generalization capacities, including VGG16, Inception V3, ResNet-152, and ResNet-50, were assessed. A large dataset of MRI images that included both tumour and non-tumor patients was used to train and test the models. The comparison study’ findings revealed encouraging performance across all models. With exceptional training accuracy of 99.97% and testing accuracy of 96.55%, VGG16 emerged as the best performance. It had exceptional recall (95.25 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup> ), accuracy (99.95 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup> ), and a shallow false positive rate. While retaining respectable precision and recall levels, ResNet-152 and Inception V3 demonstrated competitive accuracy rates of 93.34% and 94.23%, respectively. The overall performance study showed that all models performed admirably in detecting brain cancers, with high true positive rates and low false positive rates. The results highlight the potential for these models to serve as useful instruments for early tumour identification, assisting medical practitioners in providing prompt and precise diagnosis. The research also emphasizes the clinical importance of the models' generalization ability, confirming their suitability for processing as-yet-untested MRI data. The study also recommends topics for further investigation, such as improving model architectures, adding multimodal imaging data, and carrying out rigorous clinical validation. The robustness and practical application of the models might be further improved by working with radiologists and other healthcare professionals. These deep learning-based tumour prediction models ultimately represent a promising development in medical imaging, opening the door for enhanced patient care and brain tumour detection.
10
Malware Analysis in IoT & Android Systems with Defensive Mechanism
Chandra Shekhar Yadav, Jagendra Singh, Aruna Yadav et al. · Electronics · 2022 · 100 citations · Full text
Learning based Driver Drowsiness Detection Model
Jagendra Singh · 2020 · 69 citations