Multimodal brain tumor detection and classification using deep saliency map and improved dragonfly optimization algorithm

International Journal of Imaging Systems and Technology, Год журнала: 2022, Номер 33(2), С. 572 - 587

Опубликована: Дек. 1, 2022

Abstract In the last decade, there has been a significant increase in medical cases involving brain tumors. Brain tumor is tenth most common type of tumor, affecting millions people. However, if it detected early, cure rate can increase. Computer vision researchers are working to develop sophisticated techniques for detecting and classifying MRI scans primarily used analysis. We proposed an automated system detection classification using saliency map deep learning feature optimization this paper. The framework was implemented stages. initial phase framework, fusion‐based contrast enhancement technique proposed. following phase, segmentation based on maps proposed, which then mapped original images active contour. Following that, pre‐trained CNN model named EfficientNetB0 fine‐tuned trained two ways: enhanced localization images. Deep transfer train both models, features extracted from average pooling layer. fused improved fusion approach known as Entropy Serial Fusion. best chosen final step dragonfly algorithm. Finally, classified extreme machine (ELM). experimental process conducted three publically available datasets achieved accuracy 95.14, 94.89, 95.94%, respectively. comparison with several neural nets shows improvement framework.

Язык: Английский

---

MobileNetV1-Based Deep Learning Model for Accurate Brain Tumor Classification

Mesopotamian Journal of Computer Science, Год журнала: 2023, Номер unknown, С. 32 - 41

Опубликована: Март 8, 2023

Brain tumors are among the most dangerous diseases that lead to mortality after a period of time from injury. Therefore, physicians and healthcare professionals advised make an early diagnosis brain follow their instructions. Magnetic resonance imaging (MRI) is operated provide sufficient practical data in detecting tumors. Applications based on artificial intelligence contribute very large role disease detection, incredible accuracy assist creating right decisions. In particular, deep learning models, which significant part intelligence, have ability diagnose process medical image datasets. this concern, one techniques (MobileNetV1model) utilized detect 1265 images gathered Kaggle platform. The behavior model studied through four main metrics. This article deduced has effect diagnosing these important metric, accuracy, as it gained result more than 97%, excellent effect.

Язык: Английский

---

BrainNet: a fusion assisted novel optimal framework of residual blocks and stacked autoencoders for multimodal brain tumor classification

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Март 11, 2024

Abstract A significant issue in computer-aided diagnosis (CAD) for medical applications is brain tumor classification. Radiologists could reliably detect tumors using machine learning algorithms without extensive surgery. However, a few important challenges arise, such as (i) the selection of most deep architecture classification (ii) an expert field who can assess output models. These difficulties motivate us to propose efficient and accurate system based on evolutionary optimization four types modalities (t1 tumor, t1ce t2 flair tumor) large-scale MRI database. Thus, CNN modified domain knowledge connected with algorithm select hyperparameters. In parallel, Stack Encoder–Decoder network designed ten convolutional layers. The features both models are extracted optimized improved version Grey Wolf updated criteria Jaya algorithm. speeds up process improves accuracy. Finally, selected fused novel parallel pooling approach that classified neural networks. Two datasets, BraTS2020 BraTS2021, have been employed experimental tasks obtained average accuracy 98% maximum single-classifier 99%. Comparison also conducted several classifiers, techniques, nets; proposed method achieved performance.

Язык: Английский

---

PoxNet22: A Fine-Tuned Model for the Classification of Monkeypox Disease Using Transfer Learning

IEEE Access, Год журнала: 2023, Номер 11, С. 24053 - 24076

Опубликована: Янв. 1, 2023

Officials in the field of public health are concerned about a new monkeypox outbreak, even though world is now experiencing an epidemic COVID-19. Similar to variola, cowpox, and vaccinia, caused by orthopoxvirus that has two strands double-stranded. The present pandemic been propagated sexually on massive scale, particularly among individuals who identify as gay or bisexual. In this particular instance, speed with which was diagnosed single most important aspect. It possible technology machine learning could be significant assistance accurately diagnosing sickness before it can spread more people. This study's goal determine solution problem developing model for diagnosis through application image processing methods. To accomplish this, data augmentation approaches have applied avoid chances model's overfitting, then transfer-learning strategy utilized apply preprocessed dataset total six different Deep Learning (DL) models. best precision, recall, accuracy performance matrices were selected after those three metrics compared one another. A called "PoxNet22" proposed performing fine-tuning performed best. PoxNet22 outperforms other methods its classification monkeypox, does 100% accuracy. outcomes study will prove extremely helpful clinicians process classifying sickness.

Язык: Английский

---

A Lightweight Deep Learning Based Microwave Brain Image Network Model for Brain Tumor Classification Using Reconstructed Microwave Brain (RMB) Images

Biosensors, Год журнала: 2023, Номер 13(2), С. 238 - 238

Опубликована: Фев. 7, 2023

Computerized brain tumor classification from the reconstructed microwave (RMB) images is important for examination and observation of development disease. In this paper, an eight-layered lightweight classifier model called image network (MBINet) using a self-organized operational neural (Self-ONN) proposed to classify into six classes. Initially, experimental antenna sensor-based imaging (SMBI) system was implemented, RMB were collected create dataset. It consists total 1320 images: 300 non-tumor, 215 each single malignant benign tumor, 200 double 190 Then, resizing normalization techniques used preprocessing. Thereafter, augmentation applied dataset make 13,200 training per fold 5-fold cross-validation. The MBINet trained achieved accuracy, precision, recall, F1-score, specificity 96.97%, 96.93%, 96.85%, 96.83%, 97.95%, respectively, six-class original images. compared with four Self-ONNs, two vanilla CNNs, ResNet50, ResNet101, DenseNet201 pre-trained models, showed better outcomes (almost 98%). Therefore, can be reliably classifying tumor(s) in SMBI system.

Язык: Английский

---

Predicting Breast Cancer Leveraging Supervised Machine Learning Techniques

Computational and Mathematical Methods in Medicine, Год журнала: 2022, Номер 2022, С. 1 - 13

Опубликована: Авг. 16, 2022

Breast cancer is one of the leading causes increasing deaths in women worldwide. The complex nature (microcalcification and masses) breast cells makes it quite difficult for radiologists to diagnose properly. Subsequently, various computer-aided diagnosis (CAD) systems have previously been developed are being used aid cells. However, due intrinsic risks associated with delayed and/or incorrect diagnosis, indispensable improve diagnostic systems. In this regard, machine learning has recently playing a potential role early precise detection cancer. This paper presents new learning-based framework that utilizes Random Forest, Gradient Boosting, Support Vector Machine, Artificial Neural Network, Multilayer Perception approaches efficiently predict from patient data. For purpose, Wisconsin Diagnostic Cancer (WDBC) dataset utilized classified using hybrid Perceptron Model (MLP) 5-fold cross-validation as working prototype. improved classification, connection-based feature selection technique also eliminates recursive features. proposed validated on two separate datasets, i.e., Prognostic (WPBC) Original (WOBC) datasets. results demonstrate accuracy 99.12% efficient data preprocessing applied input

Язык: Английский

---

Brain tumor classification and detection via hybrid alexnet-gru based on deep learning

Biomedical Signal Processing and Control, Год журнала: 2023, Номер 89, С. 105716 - 105716

Опубликована: Ноя. 17, 2023

Язык: Английский

---

Optimized deep learning model for comprehensive medical image analysis across multiple modalities

Neurocomputing, Год журнала: 2024, Номер 619, С. 129182 - 129182

Опубликована: Дек. 12, 2024

Язык: Английский

---

Deep Learning Models Performance Evaluations for Remote Sensed Image Classification

IEEE Access, Год журнала: 2022, Номер 10, С. 111784 - 111793

Опубликована: Янв. 1, 2022

Deep learning-based land cover and use (LCLU) classification systems are a significant aspiration for remote sensing communities. In nature, images have various properties that need to be analyzed. Analyzing interpreting image is difficult due the nature of image, sensor technology's capability, other determinant variables such as seasons weather conditions. The problem essential environmental monitoring, agricultural decision-making, urban planning if it can supported by deep learning systems. Therefore, approaches proposed quickly analyze interpret classify LCLU. methods could designed starting from scratch or using pre-trained networks. However, there few comparisons developed trained on Thus, we evaluating comparing models convolutional neural network feature extractor (CNN-FE) developing scratch, transfer learning, fine-tuning LCLU system sensed images. Using CNN-FE, TL, examples, this paper compares analyzes algorithms classification. After training each model UCM dataset, evaluated compared their performances performance measurement metrics accuracy, precision, recall, f1-score, confusion matrix. adapt learn features images, TL significantly improved. As result efficient time used models, discovered fine-tuned achieved profound accuracy results in dataset.

Язык: Английский

---

TumorGANet: A Transfer Learning and Generative Adversarial Network- Based Data Augmentation Model for Brain Tumor Classification

IEEE Access, Год журнала: 2024, Номер 12, С. 103060 - 103081

Опубликована: Янв. 1, 2024

Diagnosing brain tumors using magnetic resonance imaging (MRI) presents significant challenges due to the complexities of segmentation and variability in tumor characteristics. To address limitations inherent traditional methods, this research employs an advanced deep learning approach, integrating ResNet50 for feature extraction Generative Adversarial Networks (GANs) data augmentation. A comprehensive evaluation ten transfer algorithms, including GoogLeNet VGG-16, was conducted classification tumors. Model performance assessed precision, recall, F1-score metrics, complemented by additional metrics such as Hamming loss Matthews correlation coefficient provide a more insight. ensure transparency image predictions, Explainable AI techniques, specifically Local Interpretable Model-Agnostic Explanations (LIME), were utilized. The study involved analysis 7023 MRI images, with TumorGANet being trained on dataset encompassing gliomas, meningiomas, non-tumorous cases, pituitary results demonstrate exceptional proposed model named TumorGANet, achieving accuracy 99.53%, precision recall rates 100%, F1 scores 99%, 0.2%.

Язык: Английский

---