A Hybrid Artistic Model Using Deepy-Dream Model and Multiple Convolutional Neural Networks Architectures

IEEE Access, Год журнала: 2023, Номер 11, С. 101443 - 101459

Опубликована: Янв. 1, 2023

The significant increase in drug abuse cases prompts developers to investigate techniques that mimic the hallucinations imagined by addicts and abusers, addition increasing demand for use of decorative images resulting from computer technologies. This research uses Deep Dream Neural Style Transfer technologies solve this problem. Despite significance researches on technology, there are several limitations existing studies, including image quality evaluation metrics. We have successfully addressed these issues improving diversifying types generated images. enhancement allows more effective simulating hallucinated Moreover, high-quality can be saved dataset enlargement, like augmentation process. Our proposed deepy-dream model combines features five convolutional neural network architectures: VGG16, VGG19, Inception v3, Inception-ResNet-v2, Xception. Additionally, we generate implementing each architecture as a separate model. employed autoencoder another method. To evaluate performance our models, utilize normalized cross-correlation structural similarity indexes values obtained those two measures 0.1863 0.0856, respectively, indicating performance. When considering content image, metrics yield 0.8119 0.3097, respectively. Whiefor style corresponding measure 0.0007 0.0073,

Язык: Английский

---

Haemorrhage diagnosis in colour fundus images using a fast-convolutional neural network based on a modified U-Net

Network Computation in Neural Systems, Год журнала: 2024, Номер unknown, С. 1 - 22

Опубликована: Фев. 12, 2024

Retinal haemorrhage stands as an early indicator of diabetic retinopathy, necessitating accurate detection for timely diagnosis. Addressing this need, study proposes enhanced machine-based diagnostic test retinopathy through updated UNet framework, adept at scrutinizing fundus images signs retinal haemorrhages. The customized underwent GPU training using the IDRiD database, validated against publicly available DIARETDB1 and datasets. Emphasizing complexity segmentation, employed preprocessing techniques, augmenting image quality data integrity. Subsequently, trained neural network showcased a remarkable performance boost, accurately identifying regions with 80% sensitivity, 99.6% specificity, 98.6% accuracy. experimental findings solidify network's reliability, showcasing potential to alleviate ophthalmologists' workload significantly. Notably, achieving Intersection over Union (IoU) 76.61% Dice coefficient 86.51% underscores system's competence. study's outcomes signify substantial enhancements in diagnosing critical conditions, promising profound improvements accuracy efficiency, thereby marking significant advancement automated retinopathy.

Язык: Английский

---

Medical image annotation and classification employing pyramidal feature specific lightweight deep convolution neural network

Computer Methods in Biomechanics and Biomedical Engineering Imaging & Visualization, Год журнала: 2023, Номер 11(5), С. 1678 - 1689

Опубликована: Фев. 20, 2023

Medical image annotation has significant potential to detect multiple tags. To specific tags and labels, most of the conventional learning algorithms took difficulty in matching tag with corresponding region medical images. Hence, annotations fail reproduce discrimination between various classes. In this research work tackle discrimination, we proposed a lightweight pyramidal feature-specific deep network. The Lightweight Convolutional Neural Network (LDCNN) architecture for classifying local visual regions annotate classified region. By employing learning, LDCNN align each its colour conversion makes low computation complexity, since exhibits degeneration by absorption. interpretability classification effectiveness increase. evaluate accuracy, compare AlexNet EfficientNet on benchmark datasets like MS-COCO, LC25000 multiclass Kather datasets. Empirical experimental performance index obtained outperforms baseline convolutional neural network architecture. achieves 99.6% 98.4% sensitivity, 97.9% specificity 99.1% F1 score. there is steady improvement efforts our

Язык: Английский

---

Optimizing Machine Learning Algorithms for Heart Disease Classification and Prediction

International Journal of Online and Biomedical Engineering (iJOE), Год журнала: 2023, Номер 19(15), С. 61 - 76

Опубликована: Окт. 25, 2023

According to the World Health Organization (WHO), cardiovascular disease is one of leading causes death worldwide. Thus, prevention this kind illness considered as a huge human health challenge. Additionally, diagnostic process often involves combination clinical examination, laboratory tests, and other procedures, which can be complex time-consuming. However, advances in medical technology research have led improved methods for diagnosing heart disease, help improve patient outcomes. Furthermore, Machine Learning (ML) shown promise helping diagnosis disease. Each method requires specific parameters produce good results. In paper, we propose support system based on optimized algorithms, Artificial Neural Network (ANN), Support Vector (SVM), K_Nearest Neighbour (KNN), Naive Bayes (NB), Decision Tree (DT) analyze major risk factors, such age, gender, high blood pressure, etc. To train validate ML models, dataset 558 patients with atherosclerosis used. work, achieved 96.67% promising accuracy level prediction ANN.

Язык: Английский

---

Recent advances in artificial intelligence-assisted endocrinology and diabetes

Опубликована: Ноя. 23, 2023

Artificial intelligence (AI) has gained attention for various reasons in recent years, surrounded by speculation, concerns, and expectations. Despite being developed since 1960, its widespread application took several decades due to limited computing power. Today, engineers continually improve system capabilities, enabling AI handle more complex tasks. Fields like diagnostics biology benefit from AI’s expansion, as the data they deal with requires sophisticated analysis beyond human capacity. This review showcases integration endocrinology, covering molecular phenotypic patient data. These examples demonstrate potential power research medicine.

Язык: Английский

---