Brain Tumor Detection and Prediction in MRI Images Utilizing a Fine-Tuned Transfer Learning Model Integrated Within Deep Learning Frameworks DOI Creative Commons
Deependra Rastogi, Prashant Johri, Massimo Donelli

et al.

Life, Journal Year: 2025, Volume and Issue: 15(3), P. 327 - 327

Published: Feb. 20, 2025

Brain tumor diagnosis is a complex task due to the intricate anatomy of brain and heterogeneity tumors. While magnetic resonance imaging (MRI) commonly used for imaging, accurately detecting tumors remains challenging. This study aims enhance classification via deep transfer learning architectures using fine-tuned learning, an advanced approach within artificial intelligence. Deep methods facilitate analysis high-dimensional MRI data, automating feature extraction process crucial precise diagnoses. In this research, several models, including InceptionResNetV2, VGG19, Xception, MobileNetV2, were employed improve accuracy detection. The dataset, sourced from Kaggle, contains non-tumor images. To mitigate class imbalance, image augmentation techniques applied. models pre-trained on extensive datasets recognize specific features in images, allowing improved versus experimental results show that Xception model outperformed other architectures, achieving 96.11%. result underscores its capability high-precision concludes particularly significantly efficiency diagnosis. These findings demonstrate potential AI support clinical decision making, leading more reliable diagnoses patient outcomes.

Language: Английский

High‐Performance Computing‐Based Brain Tumor Detection Using Parallel Quantum Dilated Convolutional Neural Network DOI
Swati Shinde, Aparna Pande

NMR in Biomedicine, Journal Year: 2025, Volume and Issue: 38(6)

Published: April 21, 2025

ABSTRACT In the healthcare field, brain tumor causes irregular development of cells in brain. One popular ways to identify and its progression is magnetic resonance imaging (MRI). However, existing methods often suffer from high computational complexity, noise interference, limited accuracy, which affect early diagnosis tumor. For resolving such issues, a high‐performance computing model, as big data‐based detection, utilized. As result, this work proposes novel approach named parallel quantum dilated convolutional neural network (PQDCNN)‐based detection using Map‐Reducer. The data partitioning prime process, done Fuzzy local information C‐means clustering (FLICM). partitioned subjected map reducer. mapper, Medav filtering removes noise, area segmentation by transformer model TransBTSV2. After segmenting part, image augmentation feature extraction are done. reducer phase, detected proposed PQDCNN. Furthermore, efficiency PQDCNN validated sensitivity, specificity metrics, ideal values 91.52%, 91.69%, 92.26% achieved.

Language: Английский

Citations

0
Brain Tumor Detection and Prediction in MRI Images Utilizing a Fine-Tuned Transfer Learning Model Integrated Within Deep Learning Frameworks DOI Creative Commons
Deependra Rastogi, Prashant Johri, Massimo Donelli

et al.

Life, Journal Year: 2025, Volume and Issue: 15(3), P. 327 - 327

Published: Feb. 20, 2025

Brain tumor diagnosis is a complex task due to the intricate anatomy of brain and heterogeneity tumors. While magnetic resonance imaging (MRI) commonly used for imaging, accurately detecting tumors remains challenging. This study aims enhance classification via deep transfer learning architectures using fine-tuned learning, an advanced approach within artificial intelligence. Deep methods facilitate analysis high-dimensional MRI data, automating feature extraction process crucial precise diagnoses. In this research, several models, including InceptionResNetV2, VGG19, Xception, MobileNetV2, were employed improve accuracy detection. The dataset, sourced from Kaggle, contains non-tumor images. To mitigate class imbalance, image augmentation techniques applied. models pre-trained on extensive datasets recognize specific features in images, allowing improved versus experimental results show that Xception model outperformed other architectures, achieving 96.11%. result underscores its capability high-precision concludes particularly significantly efficiency diagnosis. These findings demonstrate potential AI support clinical decision making, leading more reliable diagnoses patient outcomes.

Language: Английский

Citations

0