Brain Tumor Classification from MRI Using Image Enhancement and Convolutional Neural Network Techniques

Brain Sciences, Journal Year: 2023, Volume and Issue: 13(9), P. 1320 - 1320

Published: Sept. 14, 2023

The independent detection and classification of brain malignancies using magnetic resonance imaging (MRI) can present challenges the potential for error due to intricate nature time-consuming process involved. complexity tumor identification primarily stems from need a comprehensive evaluation spanning multiple modules. advancement deep learning (DL) has facilitated emergence automated medical image processing diagnostics solutions, thereby offering resolution this issue. Convolutional neural networks (CNNs) represent prominent methodology in visual categorization. study introduces novel integrating enhancement techniques, specifically, Gaussian-blur-based sharpening Adaptive Histogram Equalization CLAHE, with proposed model. This approach aims effectively classify different categories tumors, including glioma, meningioma, pituitary tumor, as well cases without tumors. algorithm underwent testing benchmarked data published literature, results were compared pre-trained models, VGG16, ResNet50, VGG19, InceptionV3, MobileNetV2. experimental findings method demonstrated noteworthy accuracy 97.84%, precision success rate 97.85%, recall an F1-score 97.90%. presented showcase exceptional accurately classifying most commonly occurring types. technique exhibited commendable generalization properties, rendering it valuable asset medicine aiding physicians making precise proficient diagnoses.

Language: Английский

---

Brain tumor classification using fine-tuned transfer learning models on magnetic resonance imaging (MRI) images

Digital Health, Journal Year: 2024, Volume and Issue: 10

Published: Jan. 1, 2024

Objective Brain tumors are a leading global cause of mortality, often to reduced life expectancy and challenging recovery. Early detection significantly improves survival rates. This paper introduces an efficient deep learning model expedite brain tumor through timely accurate identification using magnetic resonance imaging images. Methods Our approach leverages transfer with six algorithms: VGG16, ResNet50, MobileNetV2, DenseNet201, EfficientNetB3, InceptionV3. We optimize data preprocessing, upsample augmentation, train the models two optimizers: Adam AdaMax. perform three experiments binary multi-class datasets, fine-tuning parameters reduce overfitting. Model effectiveness is analyzed various performance scores without cross-validation. Results With smaller achieve 100% accuracy in both training testing After applying cross-validation, framework records outstanding 99.96% receiver operating characteristic on average across five tests. For larger ranges from 96.34% 98.20% different models. The methodology also demonstrates small computation time, contributing its reliability speed. Conclusion study establishes new standard for classification, surpassing existing methods efficiency. approach, incorporating advanced algorithms optimized processing, provides robust rapid solution detection.

Language: Английский

---

New Cognitive Computational Strategy for Optimizing Brain Tumour Classification using Magnetic Resonance Imaging Data

Intelligence-Based Medicine, Journal Year: 2025, Volume and Issue: unknown, P. 100215 - 100215

Published: Feb. 1, 2025

Language: Английский

---

Brain tumor detection and classification in MRI using hybrid ViT and GRU model with explainable AI in Southern Bangladesh

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Oct. 1, 2024

Language: Английский

---

Feature Extraction Using Hybrid Approach of VGG19 and GLCM For Optimized Brain Tumor Classification

International Journal of Computational and Experimental Science and Engineering, Journal Year: 2024, Volume and Issue: 10(4)

Published: Dec. 25, 2024

A brain tumor is among the illnesses that are fatal. This rationale behind significance of early disease detection. Intelligent techniques always needed to assist researchers and medical professionals in diagnosing tumors. Today's doctors employ a variety approaches identify illness. The most popular technique involves getting an MRI analyzing it look for specific diseases. However, manually evaluating pictures quite complex time-consuming. As result, attempts made discover novel methods cutting down on prediction time. Deep learning algorithms spotting tumor. Many deep employed, including CNN, RNN, LSTM, others. There benefits drawbacks related these methods. One widely utilized categorization CNN. It's critical best features while classifying Resnet, AlexNet, VGGNet, DenseNet some feature extraction employed. In this research, we proposed method extracts unique high-quality using hybrid approach VGG19 GLCM. CNN then used classify resulting images. suggested method's performance evaluation metrics—specificity, sensitivity, ROC, accuracy, loss—are examined. yields 0.98 accuracy. algorithm's sensitivity specificity 0.97 0.99, respectively. model examined by contrasting with currently use.

Language: Английский

---

Explainable ensemble deep learning-based model for brain tumor detection and classification

Neural Computing and Applications, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 20, 2024

Abstract Brain tumors are very dangerous as they cause death. A lot of people die every year because brain tumors. Therefore, accurate classification and detection in the early stages can help recovery. Various deep learning techniques have achieved good results tumor classification. The traditional methods training neural network from scratch time-consuming last for weeks training. this work, we proposed an ensemble approach depending on transfer that utilizes pre-trained models DenseNet121 InceptionV3 to detect three forms tumors: meningioma, glioma, pituitary. While developing model, some changes were made architecture by replacing their classifiers (fully connected SoftMax layers) with a new classifier adopt recent task. In addition, gradient-weighted class activation maps (Grad-CAM) explainable model verify achieve high confidence. suggested was validated using publicly available dataset 99.02% accuracy, 98.75% precision, 98.98% recall, 98.86% F1 score. outperformed others detecting classifying MRI data, verifying degree trust.

Language: Английский

---

Improved Multiclass Brain Tumor Detection via Customized Pretrained EfficientNetB7 Model

IEEE Access, Journal Year: 2023, Volume and Issue: 11, P. 117210 - 117230

Published: Jan. 1, 2023

A brain tumor considered the deadliest disease in world. Patients with misdiagnoses and insufficient treatment have a lower chance of surviving for life. However, diagnosing disturbance brain, magnetic resonance images play vital role, but due to large number produced by MRI, it is time-consuming difficult diagnose patient at an earlier stage high accuracy. Brain detection challenging variations its size, location, similarity health tissues. Further, early tumors plays role enabling wider range options, improvements survival rate, patients quality life reduced healthcare costs. That's why there need automatic, accurate system that can detect result. So, this research article, pretrained modals such as AlexNet, VGG16, VGG19, ResNet50, InceptionV3, DenseNet121, all variants EfficientNet model were chosen because their exceptional performance tasks like feature extraction image classification capability detecting anomalies images. The experiment was done on publicly available multiclass dataset tumor. before passing model, preprocessing applied terms resizing same size; further cropped extra boundaries around images; noise removed from using FastNIMeans Denoising colored filter; data augmentation technique reduce overfitting problem ensure fast training. On different deep learning models, EfficientNetB7 better results validation Further customized adding few layers fine-tuning parameters. Our (CPEB7) evaluated accuracy, loss, precision, sensitivity, specificity, recall, f1-score, MIOU (mean intersection over union) achieved accuracy 98.57%, 98.97%, 99.38% no, meningioma, pituitary, glioma classes, respectively. our proposed overall 98.97% 1.02% miss rate 95.73%. also k-fold cross-validation 99.097% fold-5. This shows superiority already existing methods evaluation

Language: Английский

---

Advances in the Use of Deep Learning for the Analysis of Magnetic Resonance Image in Neuro-Oncology

Cancers, Journal Year: 2024, Volume and Issue: 16(2), P. 300 - 300

Published: Jan. 10, 2024

Machine Learning is entering a phase of maturity, but its medical applications still lag behind in terms practical use. The field oncological radiology (and neuro-oncology particular) at the forefront these developments, now boosted by success Deep-Learning methods for analysis images. This paper reviews detail some most recent advances use Deep this field, from broader topic development Machine-Learning-based analytical pipelines to specific instantiations neuro-oncology; latter including groundbreaking ultra-low magnetic resonance imaging.

Language: Английский

---

Brain Tumor Classification From Magnetic Resonance Imaging Using Deep Learning and Novel Data Augmentation

Advances in business strategy and competitive advantage book series, Journal Year: 2023, Volume and Issue: unknown, P. 392 - 413

Published: Oct. 16, 2023

The complex and time-consuming nature of magnetic resonance imaging (MRI) may make it difficult to autonomously diagnose tumors in the brain, possibly leading erroneous detection classification. Identifying brain is a process due primarily relying on multiple modules for comprehensive evaluation. In response, advances deep learning have paved way automated medical image analysis diagnostics. Convolutional neural networks (CNNs) are crucial visual current investigation presents novel approach data augmentation that integrated with state-of-the-art models, namely Efficient-NetB0, VGG16, ResNet50, InceptionV3, MobileNetV2, accurately classify various types tumors, including glioma, meningioma, pituitary tumors. algorithm was subjected testing utilizing benchmark from existing literature.

Language: Английский

---

Handwriting-Based Detection of Schizophrenia and Bipolar Disorder Using Machine Learning

IEEE Sensors Journal, Journal Year: 2025, Volume and Issue: 25(5), P. 9113 - 9120

Published: Jan. 20, 2025

Language: Английский

---