MFEUsLNet: Skin cancer detection and classification using integrated AI with multilevel feature extraction-based unsupervised learning

Engineering Science and Technology an International Journal, Journal Year: 2024, Volume and Issue: 51, P. 101632 - 101632

Published: Feb. 7, 2024

Skin Cancer is the most common form of disease and responsible for millions deaths each year. Most relevant studies concentrate on algorithms that are based machine learning, few deep learning as well. However, due to several challenges in dermoscopic image acquisition, these unable deliver highest possible level accuracy specificity. Therefore, this article implements skin cancer detection classification (SCDC) system using multilevel feature extraction (MFE)-based artificial intelligence (AI) with unsupervised (USL), here after denoted MFEUsLNet. Initially, given images preprocessed bilateral filter, which removes noise artifacts from source images. Then, a well-known USL approach named K-means clustering (KMC) used segmentation lesion, can detect affected lesion quite efficiently. gray co-occurrence matrix (GLCM), redundant discrete wavelet transform (RDWT) low level, texture colour extraction. Finally, recurrent neural network (RNN) classifier train multi-level features classify multiple types cancer. The simulations proven proposed MFEUsLNet model outperformed state-of-the-art SCDC approaches terms medical statistical quality metrics such accuracy, specificity, precision, recall, F1-score, sensitivity ISIC-2020 dataset.

Language: Английский

---

A machine learning and deep learning-based integrated multi-omics technique for leukemia prediction

Heliyon, Journal Year: 2024, Volume and Issue: 10(3), P. e25369 - e25369

Published: Feb. 1, 2024

In recent years, scientific data on cancer has expanded, providing potential for a better understanding of malignancies and improved tailored care. Advances in Artificial Intelligence (AI) processing power algorithmic development position Machine Learning (ML) Deep (DL) as crucial players predicting Leukemia, blood cancer, using integrated multi-omics technology. However, realizing these goals demands novel approaches to harness this deluge. This study introduces Leukemia diagnosis approach, analyzing accuracy ML DL algorithms. techniques, including Random Forest (RF), Naive Bayes (NB), Decision Tree (DT), Logistic Regression (LR), Gradient Boosting (GB), methods such Recurrent Neural Networks (RNN) Feedforward (FNN) are compared. GB achieved 97 % ML, while RNN outperformed by achieving 98 DL. approach filters unclassified effectively, demonstrating the significance leukemia prediction. The testing validation was based 17 different features patient age, sex, mutation type, treatment methods, chromosomes, others. Our compares techniques chooses best technique that gives optimum results. emphasizes implications high-throughput technology healthcare, offering

Language: Английский

---

Dataset and Feature Analysis for Diabetes Mellitus Classification using Random Forest

Journal of Computing Theories and Applications, Journal Year: 2023, Volume and Issue: 1(1), P. 41 - 48

Published: Sept. 28, 2023

Diabetes Mellitus is a hazardous disease, and according to the World Health Organization (WHO), diabetes will be one of main causes death by 2030. One most popular datasets PIMA Indians, this dataset has been widely tested on various machine learning (ML) methods, even deep (DL). But average, ML methods are not able produce good accuracy. The quality features influential thing in case, so deeper investment needed examine dataset. This research analyze compare Indians Abelvikas using Random Forest (RF) method. two imbalanced, fact, more imbalanced larger number classes it complex. RF was chosen because that best results datasets. Based test results, very contrasting were obtained had accuracy, precision, recall, reaching 100%, only achieved 75% for 87% 80% recall. Testing done with 3, 5, 7, 10, 15 tree parameters. Apart from that, also k-fold validation get valid results. determines much better complete glucose support them.

Language: Английский

---

RICE DISEASE RECOGNITION USING TRANSFER LEARNING XCEPTION CONVOLUTIONAL NEURAL NETWORK

Jurnal Teknik Informatika (Jutif), Journal Year: 2023, Volume and Issue: 4(6), P. 1535 - 1540

Published: Dec. 26, 2023

As one of the major rice producers, Indonesia faces significant challenges related to plant diseases such as blast, brown spot, tugro, leaf smut, and blight. These threaten food security result in economic losses, underscoring importance early detection management diseases. Convolutional Neural Network (CNN) has proven effective detecting plants. Specifically, transfer learning with CNN, particularly Xception model, advantage efficiently extracting automatic features performing well even limited datasets. This study aims develop model for disease recognition based on images. Through fine-tuning process, achieved accuracies, precisions, recalls, F1-scores 0.89, 0.90, respectively, a dataset total 320 Additionally, outperformed VGG16, MobileNetV2, EfficientNetV2.

Language: Английский

---

AI-driven health analysis for emerging respiratory diseases: A case study of Yemen patients using COVID-19 data

Mathematical Biosciences & Engineering, Journal Year: 2025, Volume and Issue: 22(3), P. 554 - 584

Published: Jan. 1, 2025

In low-income and resource-limited countries, distinguishing COVID-19 from other respiratory diseases is challenging due to similar symptoms the prevalence of comorbidities. Yemen, acute comorbidities further complicate differentiation between infectious diseases. We explored use AI-powered predictive models classifiers enhance healthcare preparedness by forecasting disease trends using data. developed mathematical based on autoregressive (AR), moving average (MA), ARMA, machine deep learning algorithms predict daily confirmed deaths. Statistical were trained 80% data tested remaining 20%, with predicted results compared actual values. The ARMA model demonstrated promising performance. Additionally, eight (ML) (DL) utilized identify severity indicators. Among ML classifiers, Decision Tree (DT) achieved highest accuracy at 74.70%, followed closely Random Forest (RF) 74.66%. DL showed comparable scores, around 70%. terms AUC-ROC, kernel Support Vector Machine (SVM) outperformed others, achieving 71% accuracy, precision, recall, F-measure, area under curve values 0.7, 0.75, 0.59, 0.72, respectively. These findings underscore potential AI-driven health analysis optimize resource allocation for

Language: Английский

---

Construction and optimization of vending machine decision support system based on improved C4.5 decision tree

Heliyon, Journal Year: 2024, Volume and Issue: 10(3), P. e25024 - e25024

Published: Jan. 23, 2024

The intensification of market competition makes refined operation management become the focus attention major manufacturers. As an important branch artificial intelligence (AI), machine learning (ML) plays a key role in it, and has its application prospect various systems. Based on this situation, paper takes vending machines as research object. On one hand, product classification model is constructed based decision tree algorithm. other neural network (NN), sales forecast built. Finally, above research, theoretical framework support system (DSS) for constructed. shows that: (1) accuracy C4.5 algorithm can reach 87 % at highest 68 lowest. improved 67 lowest, with little difference between them. (2) maximum running time about 5500 ms, minimum close to 1 ms. In addition, all seven datasets better than that unmodified (3) When back propagation (BPNN) used machines, curve predicted data basically coincides actual data, which high. This aims build convenient secure DSS by taking example. also uses reinforcement optimize methods paper. It further performance efficiency provide service experience customers. Meanwhile, use make whole more intelligent adaptive cope changing environment.

Language: Английский

---

Transformer-based decoder of melanoma classification using hand-crafted texture feature fusion and Gray Wolf Optimization algorithm

MethodsX, Journal Year: 2024, Volume and Issue: 13, P. 102839 - 102839

Published: July 3, 2024

Melanoma is a type of skin cancer that poses significant health risks and requires early detection for effective treatment. This study proposing novel approach integrates transformer-based model with hand-crafted texture features Gray Wolf Optimization, aiming to enhance efficiency melanoma classification. Preprocessing involves standardizing image dimensions enhancing quality through median filtering techniques. Texture features, including GLCM LBP, are extracted capture spatial patterns indicative melanoma. The GWO algorithm applied select the most discriminative features. A decoder then employed classification, leveraging attention mechanisms contextual dependencies. experimental validation on HAM10000 dataset ISIC2019 showcases effectiveness proposed methodology. model, integrated guided by achieves outstanding results. results showed method performed well in tasks, achieving an accuracy F1-score 99.54% 99.11% dataset, 99.47%, 99.25% dataset. • We use concepts LBP extract from lesion images. Optimization (GWO) feature selection. based Transformers utilized

Language: Английский

---

Breaking Boundaries in Diagnosis: Non-Invasive Anemia Detection Empowered by AI

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 9292 - 9307

Published: Jan. 1, 2024

This article evolved because several instances of anemia are still discovered too late, especially in communities with limited medical resources and access to laboratory tests. Invasive diagnostic technologies expensive expenses additional impediments early diagnosis. To detect anemia, an effective, accurate, non-invasive method is required. In this study, the conjunctival image eye analyzed as a detecting anemia. Various model approaches were tested endeavor categorize anemic healthy patients accurately possible. The Support Vector Machine (SVM) algorithm-integrated MobileNetV2 was determined be most effective plan. With combination, accuracy 93%, sensitivity 91%, specificity 94%. These findings show that can successfully identify while identifying patients. offers means on, making it promising for use clinical settings. SVM+MobileNetV2 technique relies on images conjunctiva has potential improve healthcare by people who may have had earlier. stands out solid option efficient precise diagnosis when accuracy, sensitivity, balanced.

Language: Английский

---

AI-Based Detection Techniques for Skin Diseases: A Review of Recent Methods, Datasets, Metrics, and Challenges

Journal of Future Artificial Intelligence and Technologies, Journal Year: 2024, Volume and Issue: 1(3), P. 318 - 336

Published: Dec. 28, 2024

The identification and early treatment of skin diseases are crucial to mitigate serious health risks. growing attention on researching disease analysis stems from the transformative impact artificial intelligence (AI) in dermatology. In this systematic review, we adhered Preferred Reporting Items for Systematic Reviews Meta-Analyses (PRISMA) guidelines comprehensively assess recent approaches detection. Our study addressed four key research questions exploring methods detection, evaluation techniques employed measure effectiveness detection models, datasets utilized, challenges encountered applying machine learning deep We screened studies 2019 2023 reputable databases, including IEEE Explore, Science Direct, Google Scholar. findings revealed that CNN model outperformed other models. Additionally, our identified ISIC public dataset as most frequently used dataset. reviewed metrics such accuracy, recall, precision, sensitivity, F1 score evaluate performance. several limitations reviewed, use limited datasets, distinguishing between with similar features, related limitations. Overall, provided a comprehensive overview current state-of-the-art highlighted future directions.

Language: Английский

---

Manifold-Regularized Feature Selector for High-Resolution Aerial Photographs Categorization

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 41354 - 41363

Published: Jan. 1, 2024

Recognizing each aerial photo with high-resolution (HR) is a useful technology in image understanding. Herein, manifold-regularized feature selection (MRFS) designed to acquire discriminative perceptual features that classify HR images into different categories. Practically, human visual cognition process reflects that, scenic picture, the less visually attractive patches are highly related. Meanwhile, foreground practically unrelated other. Following this observation, we work propose multi-layer low-rank paradigm which calculates succinct set of foreground. We sequentially link above build so-called gaze shifting path (GSP). GSP can mimick how humans perceiving images. Afterward, formulate MRFS framework obtain subset high quality from entire deep representation. Thereby, an SVM learned simultaneously. Moreover, distribution on underlying manifold be maximally preserved during (FS). To comprehensively evaluate our method, collect massive-scale containing over 4.87 million high- and low-resolution Extensive empirical validations have shown algorithm's efficiency effectiveness: 1) testing time cost 0.8s faster than second best one categorize image, 2) average categorization accuracy 4.5% higher one.

Language: Английский

---