Privacy-Preserving Breast Cancer Classification: A Federated Transfer Learning Approach

Deleted Journal, Год журнала: 2024, Номер 37(4), С. 1488 - 1504

Опубликована: Фев. 29, 2024

Breast cancer is deadly causing a considerable number of fatalities among women in worldwide. To enhance patient outcomes as well survival rates, early and accurate detection crucial. Machine learning techniques, particularly deep learning, have demonstrated impressive success various image recognition tasks, including breast classification. However, the reliance on large labeled datasets poses challenges medical domain due to privacy issues data silos. This study proposes novel transfer approach integrated into federated framework solve limitations limited collaborative healthcare settings. For classification, mammography MRO images were gathered from three different centers. Federated an emerging privacy-preserving paradigm, empowers multiple institutions jointly train global model while maintaining decentralization. Our proposed methodology capitalizes power pre-trained ResNet, neural network architecture, feature extractor. By fine-tuning higher layers ResNet using diverse centers, we enable learn specialized features relevant domains leveraging comprehensive representations acquired large-scale like ImageNet. overcome shift caused by variations distributions across introduce adversarial training. The learns minimize discrepancy maximizing classification accuracy, facilitating acquisition domain-invariant features. We conducted extensive experiments obtained Comparative analysis was performed evaluate against traditional standalone training without adaptation. When compared with models, our showed accuracy 98.8% computational time 12.22 s. results showcase promising enhancements generalization, underscoring potential method improving performance upholding environment.

Язык: Английский

---

Advancements in Plant Leaf Disease Classification: Integrating Machine Learning and Graph Convolutional Networks for Sustainable Agriculture

Опубликована: Март 14, 2024

In order to support life on Earth and preserve ecological equilibrium, plants are essential. However, a variety of biotic abiotic elements constantly pose threat them; plant leaves especially susceptible illness. The ethology leaf diseases, their consequences health, mitigation strategies mined in this research. categorization diseases both infectious noninfectious the main topic discussion. These illnesses threaten global foo d security ecosystems because sensitivity infections, environment, genetics. significance effective disease management is emphasized article, which takes into account like sustainable practices, data driven forecasting, precision agriculture. application Artificial Intelligence (AI) Machine Learning (ML) touted as potent technique for classifying diseases. proposed approach, makes use graph convolutional networks, offers fresh way get around current problems. importance high-quality datasets creating precise decision systems covered study. It how adaptable ML models are, enabling ongoing learning identification novel illnesses. Notwithstanding progress made, obstacles including class imbalance massive problems recognized, highlighting necessity meticulous experimental planning model training.

Язык: Английский

---

Computer Vision and Creative Content Generation: Text-to-Sketch Conversion

2022 International Conference on Communication, Computing and Internet of Things (IC3IoT), Год журнала: 2024, Номер unknown, С. 1 - 6

Опубликована: Апрель 17, 2024

Within the field of computer vision and creative content generation, process combining visual elements based on textual descriptions has emerged as a captivating area study advancement. An intriguing application in this is text-to-sketch conversion, which employs advanced machine learning methods to convert written into equivalent sketches or drawings. The utilization representation consistently proven be superior method communication. Therefore, incorporating visualization any communicative domain will greatly enhance efficiency process. objective paper accomplish aim. This presents development an image generator that produces sketch picture using user's provided description. produced outline exhibited HTML canvas within website. POS tagging parse description, then utilizes hidden layer regression neural network Sketch RNN generate desired parsed model, been pre-trained Quick Draw dataset, employed draw objects mentioned Additionally, utilized position instructions. result consists series brushstrokes are rendered canvas. software can by system user computer, allowing them input their string keyboard. operate efficiently reliably, even PCs with minimal hardware specifications. A pre-installed browser allows view interface web page, where they information receive corresponding output.

Язык: Английский

---

Brain tumor classification utilizing Triple Memristor Hopfield Neural Network optimized with Northern Goshawk Optimization for MRI image

Biomedical Signal Processing and Control, Год журнала: 2024, Номер 95, С. 106450 - 106450

Опубликована: Май 23, 2024

Язык: Английский

---

Parkinson’s Disease Detection: VGG-ResNet Hybrid Approach

Опубликована: Апрель 12, 2024

This study integrates the capabilities of VGG-16 and ResNet-50 convolutional neural networks to provide a novel method for Parkinson's Disease (PD) detection. The approach uses methodical preparation strategy adapts last layers binary classification, all while leveraging carefully annotated dataset. combined model shows encouraging accuracy, highlighting advantages using both architectures in concert. robustness suggested methodology is supported by thorough assessments that include cross-validation, precision, recall, F1-score measures. field medical image analysis gains from this research, which lays groundwork automated diagnostic tools can help professionals identify disease early.

Язык: Английский

---

Enhancing Lung Nodule Classification: A Novel CViEBi-CBGWO Approach with Integrated Image Preprocessing

Deleted Journal, Год журнала: 2024, Номер 37(5), С. 2108 - 2125

Опубликована: Март 25, 2024

Cancer detection and accurate classification pose significant challenges for medical professionals, as it is described a lethal illness. Diagnosing the malignant lung nodules in its initial stage significantly enhances recovery survival rates. Therefore, novel model named convolutional vision Elman bidirectional–based crossover boosted grey wolf optimization (CViEBi-CBGWO) has been proposed to enhance accuracy. CT images selected further preprocessing are obtained from LUNA16 dataset LIDC-IDRI dataset. The data undergoes phases involving normalization, augmentation, filtering improve generalization ability well image quality. local features within preprocessed extracted by implementing neural network (CNN). For extracting global images, transformer (ViT) consists of five encoder blocks. attained combined generate feature map. bidirectional long short-term memory (EBiLSTM) applied categorize generated map benign malignant. operation integrated with (GWO) algorithm, form CBGWO fine-tunes parameters CViEBi model, eliminating problem optima. Experimental validation conducted using various evaluation measures assess effectiveness. Comparative analysis demonstrates superior accuracy 98.72% method compared existing methods.

Язык: Английский

---

Anomaly graph: leveraging dynamic graph convolutional networks for enhanced video anomaly detection in surveillance and security applications

Neural Computing and Applications, Год журнала: 2024, Номер 36(20), С. 12011 - 12028

Опубликована: Апрель 18, 2024

Язык: Английский

---

SAIF-Cnet: self-attention improved faster convolutional neural network for decentralized blockchain-based key management protocol

Wireless Networks, Год журнала: 2024, Номер 30(5), С. 3211 - 3228

Опубликована: Апрель 13, 2024

Язык: Английский

---

CitrusDiseaseNet: An integrated approach for automated citrus disease detection using deep learning and kernel extreme learning machine

Earth Science Informatics, Год журнала: 2024, Номер 17(4), С. 3053 - 3070

Опубликована: Май 21, 2024

Язык: Английский

---

Unlocking Sign Language Communication: A Deep Learning Paradigm for Overcoming Accessibility Challenges

2022 International Conference on Communication, Computing and Internet of Things (IC3IoT), Год журнала: 2024, Номер unknown, С. 1 - 6

Опубликована: Апрель 17, 2024

One of the most important modes engagement that may be utilized while communicating with people who are deaf and mute is use sign language. Despite this, there obstacles must overcome before language can become widely used. This because not everyone capable learning language, which results in communication hurdles. The goal this project to take advantage deep as a game-changing option for improving interaction among community or mute. In past, many technologies have been developed address problem relied on external sensors, has created accessibility issues. For purpose investigation, we make OpenCV acquire images deploy Convolutional Neural Network (CNN) approaches train machine. final output then transformed into text. Our research especially aimed at entire acceptance American Sign Language (ASL), consists 26 letters 10 numbers. Previous studies offered techniques partially recognition, but study specifically focused full accepted vast majority characters static, although few incorporate dynamic gestures. offers an unusual task As result, primary focus our extraction characteristics from actions hands fingers order differentiate passive active Through application strategies, all-encompassing strategy intends constraints now available pave way towards communities more accessible effective.

Язык: Английский

---