BCNet: A Deep Learning Computer-Aided Diagnosis Framework for Human Peripheral Blood Cell Identification

Diagnostics, Journal Year: 2022, Volume and Issue: 12(11), P. 2815 - 2815

Published: Nov. 16, 2022

Blood cells carry important information that can be used to represent a person's current state of health. The identification different types blood in timely and precise manner is essential cutting the infection risks people face on daily basis. BCNet an artificial intelligence (AI)-based deep learning (DL) framework was proposed based capability transfer with convolutional neural network rapidly automatically identify eight-class scenario: Basophil, Eosinophil, Erythroblast, Immature Granulocytes, Lymphocyte, Monocyte, Neutrophil, Platelet. For purpose establishing dependability viability BCNet, exhaustive experiments consisting five-fold cross-validation tests are carried out. Using strategy, we conducted in-depth comprehensive BCNet's architecture test it three optimizers ADAM, RMSprop (RMSP), stochastic gradient descent (SGD). Meanwhile, performance directly compared using same dataset state-of-the-art models DensNet, ResNet, Inception, MobileNet. When employing optimizers, demonstrated better classification ADAM RMSP optimizers. best evaluation achieved optimizer terms 98.51% accuracy 96.24% F1-score. Compared baseline model, clearly improved prediction 1.94%, 3.33%, 1.65% RMSP, SGD, respectively. model outperformed AI DenseNet, MobileNet testing time single cell image by 10.98, 4.26, 2.03, 0.21 msec. In comparison most recent models, could able generate encouraging outcomes. It for advancement healthcare facilities have such recognition rate improving detection cells.

Language: Английский

---

ETECADx: Ensemble Self-Attention Transformer Encoder for Breast Cancer Diagnosis Using Full-Field Digital X-ray Breast Images

Diagnostics, Journal Year: 2022, Volume and Issue: 13(1), P. 89 - 89

Published: Dec. 28, 2022

Early detection of breast cancer is an essential procedure to reduce the mortality rate among women. In this paper, a new AI-based computer-aided diagnosis (CAD) framework called ETECADx proposed by fusing benefits both ensemble transfer learning convolutional neural networks as well self-attention mechanism vision transformer encoder (ViT). The accurate and precious high-level deep features are generated via backbone network, while used diagnose probabilities in two approaches: Approach A (i.e., binary classification) B multi-classification). To build CAD system, benchmark public multi-class INbreast dataset used. Meanwhile, private real images collected annotated expert radiologists validate prediction performance framework. promising evaluation results achieved using mammograms with overall accuracies 98.58% 97.87% for approaches, respectively. Compared individual networks, model improves 6.6% 4.6% approaches. hybrid shows further improvement when ViT-based network 8.1% 6.2% diagnosis, For validation purposes images, system provides encouraging 97.16% 89.40% has capability predict lesions single mammogram average 0.048 s. Such could be useful helpful assist practical applications providing second supporting opinion distinguishing various malignancies.

Language: Английский

---

Classification of Breast Cancer Using Transfer Learning and Advanced Al-Biruni Earth Radius Optimization

Biomimetics, Journal Year: 2023, Volume and Issue: 8(3), P. 270 - 270

Published: June 26, 2023

Breast cancer is one of the most common cancers in women, with an estimated 287,850 new cases identified 2022. There were 43,250 female deaths attributed to this malignancy. The high death rate associated type can be reduced early detection. Nonetheless, a skilled professional always necessary manually diagnose malignancy from mammography images. Many researchers have proposed several approaches based on artificial intelligence. However, they still face obstacles, such as overlapping cancerous and noncancerous regions, extracting irrelevant features, inadequate training models. In paper, we developed novel computationally automated biological mechanism for categorizing breast cancer. Using optimization approach Advanced Al-Biruni Earth Radius (ABER) algorithm, boosting classification realized. stages framework include data augmentation, feature extraction using AlexNet transfer learning, optimized convolutional neural network (CNN). learning CNN improved accuracy when results are compared recent approaches. Two publicly available datasets utilized evaluate framework, average 97.95%. To ensure statistical significance difference between methodology, additional tests conducted, analysis variance (ANOVA) Wilcoxon, addition evaluating various metrics. these emphasized effectiveness methodology current methods.

Language: Английский

---

Improved Breast Cancer Classification through Combining Transfer Learning and Attention Mechanism

Life, Journal Year: 2023, Volume and Issue: 13(9), P. 1945 - 1945

Published: Sept. 21, 2023

Breast cancer, a leading cause of female mortality worldwide, poses significant health challenge. Recent advancements in deep learning techniques have revolutionized breast cancer pathology by enabling accurate image classification. Various imaging methods, such as mammography, CT, MRI, ultrasound, and biopsies, aid detection. Computer-assisted pathological classification is paramount importance for diagnosis. This study introduces novel approach to histopathological It leverages modified pre-trained CNN models attention mechanisms enhance model interpretability robustness, emphasizing localized features discrimination complex cases. Our method involves transfer with models—Xception, VGG16, ResNet50, MobileNet, DenseNet121—augmented the convolutional block module (CBAM). The are finetuned, two CBAM incorporated at end models. compared state-of-the-art diagnosis approaches tested accuracy, precision, recall, F1 score. confusion matrices used evaluate visualize results They help assessing models’ performance. test accuracy rates mechanism (AM) using Xception on “BreakHis” dataset encouraging 99.2% 99.5%. DenseNet121 AMs 99.6%. proposed also performed better than previous examined related studies.

Language: Английский

---

An intelligent healthcare framework for breast cancer diagnosis based on the information fusion of novel deep learning architectures and improved optimization algorithm

Engineering Applications of Artificial Intelligence, Journal Year: 2024, Volume and Issue: 137, P. 109152 - 109152

Published: Aug. 22, 2024

Language: Английский

---

CNN-Hyperparameter Optimization for Diabetic Maculopathy Diagnosis in Optical Coherence Tomography and Fundus Retinography

Mathematics, Journal Year: 2022, Volume and Issue: 10(18), P. 3274 - 3274

Published: Sept. 9, 2022

Diabetic Maculopathy (DM) is considered the most common cause of permanent visual impairment in diabetic patients. The absence clear pathological symptoms DM hinders timely diagnosis and treatment such a critical condition. Early feasible through eye screening technologies. However, manual inspection retinography images by specialists time-consuming routine. Therefore, many deep learning-based computer-aided systems have been recently developed for automatic prognosis retinal images. Manual tuning learning network’s hyperparameters practice literature. hyperparameter optimization has shown to be promising improving performance networks classifying several diseases. This study investigates impact using Bayesian (BO) algorithm on classification detecting In this research, we propose two new custom Convolutional Neural Network (CNN) models detect distinct types photography; Optical Coherence Tomography (OCT) fundus datasets. approach utilized determine optimal architectures proposed CNNs optimize their hyperparameters. findings reveal effectiveness fine-tuning model maculopathy OCT pre-trained CNN AlexNet, VGG16Net, VGG 19Net, GoogleNet, ResNet-50 are employed compared with CNN-based models. Statistical analyses, based one-way analysis variance (ANOVA) test, receiver operating characteristic (ROC) curve, histogram, performed confirm

Language: Английский

---

Multimodal breast cancer hybrid explainable computer-aided diagnosis using medical mammograms and ultrasound Images

Journal of Applied Biomedicine, Journal Year: 2024, Volume and Issue: 44(3), P. 731 - 758

Published: July 1, 2024

Language: Английский

---

Detection and Classification of Histopathological Breast Images Using a Fusion of CNN Frameworks

Diagnostics, Journal Year: 2023, Volume and Issue: 13(10), P. 1700 - 1700

Published: May 11, 2023

Breast cancer is responsible for the deaths of thousands women each year. The diagnosis breast (BC) frequently makes use several imaging techniques. On other hand, incorrect identification might occasionally result in unnecessary therapy and diagnosis. Therefore, accurate can save a significant number patients from undergoing surgery biopsy procedures. As recent developments field, performance deep learning systems used medical image processing has showed benefits. Deep (DL) models have found widespread aim extracting important features histopathologic BC images. This helped to improve classification assisted automation process. In times, both convolutional neural networks (CNNs) hybrid learning-based approaches demonstrated impressive performance. this research, three different types CNN are proposed: straightforward model (1-CNN), fusion (2-CNN), (3-CNN). findings experiment demonstrate that techniques based on 3-CNN algorithm performed best terms accuracy (90.10%), recall (89.90%), precision (89.80%), f1-Score (89.90%). conclusion, CNN-based been developed contrasted with more modern machine models. application methods resulted increase classification.

Language: Английский

---

Clinical Decision Support Framework for Segmentation and Classification of Brain Tumor MRIs Using a U-Net and DCNN Cascaded Learning Algorithm

Healthcare, Journal Year: 2022, Volume and Issue: 10(12), P. 2340 - 2340

Published: Nov. 22, 2022

Brain tumors (BTs) are an uncommon but fatal kind of cancer. Therefore, the development computer-aided diagnosis (CAD) systems for classifying brain in magnetic resonance imaging (MRI) has been subject many research papers so far. However, this sector is still its early stage. The ultimate goal to develop a lightweight effective implementation U-Net deep network use performing exact real-time segmentation. Moreover, simplified convolutional neural (DCNN) architecture BT classification presented automatic feature extraction and segmented regions interest (ROIs). Five layers, rectified linear unit, normalization, max-pooling layers make up DCNN's proposed architecture. introduced method was verified on multimodal tumor segmentation (BRATS 2015) datasets. Our experimental results BRATS 2015 acquired Dice similarity coefficient (DSC) scores, sensitivity, accuracy 88.8%, 89.4%, 88.6% high-grade gliomas. When it comes segmenting images, performance our CAD framework par with existing state-of-the-art methods. achieved study images improved upon reported prior studies. Image from 88% 88.6%.

Language: Английский

---

RNN and BiLSTM Fusion for Accurate Automatic Epileptic Seizure Diagnosis Using EEG Signals

Life, Journal Year: 2022, Volume and Issue: 12(12), P. 1946 - 1946

Published: Nov. 22, 2022

Epilepsy is a common neurological condition. The effects of epilepsy are not restricted to seizures alone. They comprise wide spectrum problems that might impair and reduce quality life. Even with medication, 30% patients still have recurring seizures. An epileptic seizure caused by significant neuronal electrical activity, which affects brain activity. EEG shows these changes as high-amplitude spiky sluggish waves. Recognizing on an electroencephalogram (EEG) manually professional neurologist time-consuming labor-intensive process, hence efficient automated approach necessary for the identification seizure. One technique increase speed accuracy diagnosis could be made utilizing computer-aided systems built deep neural networks, or DNN. This study introduces fusion recurrent networks (RNNs) bi-directional long short-term memories (BiLSTMs) automatic via signal processing in order tackle aforementioned informational challenges. electroencephalogram’s raw data were first normalized after undergoing pre-processing. A RNN model was fed sequence trained accurately extract features from data. Afterwards, passed BiLSTM layers so further temporal information retrieved. In addition, proposed RNN-BiLSTM tested experimental setting using freely accessible UCI dataset. Experimental findings suggested achieved avg values 98.90%, 98.50%, 98. 20%, 98.60%, respectively, accuracy, sensitivity, precision, specificity. To verify new model’s efficacy, it compared other models, such RNN-LSTM RNN-GRU learning shown improved same metrics 1.8%, 1.69%, 1.95%, 2.2% 5-fold. Additionally, method state-of-the-art approaches proved more accurate categorization techniques.

Language: Английский

---