Clinical Applications of Artificial Intelligence in Medical Imaging and Image Processing—A Review

Cancers, Journal Year: 2024, Volume and Issue: 16(10), P. 1870 - 1870

Published: May 14, 2024

Artificial intelligence (AI) is currently becoming a leading field in data processing [...]

Language: Английский

---

Boosting skin cancer diagnosis accuracy with ensemble approach

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Jan. 8, 2025

Skin cancer is common and deadly, hence a correct diagnosis at an early age essential. Effective therapy depends on precise classification of the several skin forms, each with special traits. Because dermoscopy other sophisticated imaging methods produce detailed lesion images, detection has been enhanced. It's still difficult to analyze images differentiate benign from malignant tumors, though. Better predictive modeling are needed since diagnostic procedures used now frequently inaccurate inconsistent results. In dermatology, Machine learning (ML) models becoming essential for automatic lesions image data. With ensemble model, which mix ML approaches take use their advantages lessen disadvantages, this work seeks improve predictions. We introduce new method, Max Voting optimization classification. On HAM10000 ISIC 2018 datasets, we trained assessed three distinct models: Random Forest (RF), Multi-layer Perceptron Neural Network (MLPN), Support Vector (SVM). Overall performance was increased by combined predictions made technique. Moreover, feature vectors that were optimally produced data Genetic Algorithm (GA) given models. demonstrate method greatly improves performance, reaching accuracy 94.70% producing best results F1-measure, recall, precision. The most dependable robust approach turned out be Voting, combines benefits numerous pre-trained provide efficient classifying lesions.

Language: Английский

---

Skin cancer identification utilizing deep learning: A survey

IET Image Processing, Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 2, 2024

Abstract Melanoma, a highly prevalent and lethal form of skin cancer, has significant impact globally. The chances recovery for melanoma patients substantially improve with early detection. Currently, deep learning (DL) methods are gaining popularity in assisting the identification melanoma. Despite their high performance, relying solely on an image classifier undermines credibility application makes it difficult to understand rationale behind model's predictions highlighting need Explainable AI (XAI). This study provides survey cancer using DL techniques utilized studies from 2017 2024. Compared existing studies, authors address latest related covering several public datasets focusing segmentation, classification based convolutional neural networks vision transformers, explainability. analysis comparisons will be beneficial researchers developers this area, identify suitable used automated classification. Thereby, findings can implement support applications advancing diagnosis process.

Language: Английский

---

AI-Driven Enhancement of Skin Cancer Diagnosis: A Two-Stage Voting Ensemble Approach Using Dermoscopic Data

Cancers, Journal Year: 2025, Volume and Issue: 17(1), P. 137 - 137

Published: Jan. 3, 2025

Background: Skin cancer is the most common worldwide, with melanoma being deadliest type, though it accounts for less than 5% of cases. Traditional skin detection methods are effective but often costly and time-consuming. Recent advances in artificial intelligence have improved diagnosis by helping dermatologists identify suspicious lesions. Methods: The study used datasets from two ethnic groups, sourced ISIC platform CSMU Hospital, to develop an AI diagnostic model. Eight pre-trained models, including convolutional neural networks vision transformers, were fine-tuned. three best-performing models combined into ensemble model, which underwent multiple random experiments ensure stability. To improve accuracy reduce false negatives, a two-stage classification strategy was employed: three-class model initial classification, followed binary secondary prediction benign Results: In dataset, negative rate malignant lesions significantly reduced, number cases misclassified as dropped 124 45. CSMUH negatives completely eliminated, reducing zero, resulting notable improvement precision reduction rate. Conclusions: Through proposed method, demonstrated clear success both datasets. First, can assist doctors distinguishing between patients who require urgent treatment, non-melanoma be treated later, that do not intervention. Subsequently, effectively reduces These findings highlight potential technology diagnosis, particularly resource-limited medical settings, where could become valuable clinical tool accuracy, mortality, healthcare costs.

Language: Английский

---

Advancing Skin Cancer Prediction Using Ensemble Models

Computers, Journal Year: 2024, Volume and Issue: 13(7), P. 157 - 157

Published: June 21, 2024

There are many different kinds of skin cancer, and an early precise diagnosis is crucial because cancer both frequent deadly. The key to effective treatment accurately classifying the various cancers, which have unique traits. Dermoscopy other advanced imaging techniques enhanced detection by providing detailed images lesions. However, interpreting these distinguish between benign malignant tumors remains a difficult task. Improved predictive modeling necessary due occurrence erroneous inconsistent outcomes in present diagnostic processes. Machine learning (ML) models become essential field dermatology for automated identification categorization lesions using image data. aim this work develop improved predictions ensemble models, combine numerous machine approaches maximize their combined strengths reduce individual shortcomings. This paper proposes fresh special approach model optimization classification: Max Voting method. We trained assessed five ISIC 2018 HAM10000 datasets: AdaBoost, CatBoost, Random Forest, Gradient Boosting, Extra Trees. Their enhance overall performance with Moreover, were fed feature vectors that optimally generated from data genetic algorithm (GA). show that, accuracy 95.80%, significantly improves when compared individually. Obtaining best results F1-measure, recall, precision, method turned out be most dependable robust. novel aspect more robustly reliably classified technique. Several pre-trained models’ benefits approach.

Language: Английский

---

Systematic Review of Deep Learning Techniques in Skin Cancer Detection

BioMedInformatics, Journal Year: 2024, Volume and Issue: 4(4), P. 2251 - 2270

Published: Nov. 14, 2024

Skin cancer is a serious health condition, as it can locally evolve into disfiguring states or metastasize to different tissues. Early detection of this disease critical because increases the effectiveness treatment, which contributes improved patient prognosis and reduced healthcare costs. Visual assessment histopathological examination are gold standards for diagnosing these types lesions. Nevertheless, processes strongly dependent on dermatologists’ experience, with excision advised only when suspected by physician. Multiple approaches have surfed over last few years, particularly those based deep learning (DL) strategies, goal assisting medical professionals in diagnosis process ultimately diminishing diagnostic uncertainty. This systematic review focused analysis relevant studies DL applications skin diagnosis. The qualitative included 164 records topic. AlexNet, ResNet-50, VGG-16, GoogLeNet architectures considered top choices obtaining best classification results, multiclassification current trend. Public databases key elements area should be maintained facilitate scientific research.

Language: Английский

---

Enhancing Arecanut Quality Grading: A Comparison of Custom CNNs and Transfer Learning Models

Research Square (Research Square), Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 20, 2025

Effective grading of arecanut is essential for ensuring product quality, maximizing market competitiveness, and satisfying consumer preferences. However, traditional methods are challenging due to variations in size, shape, appearance, resulting subjective inconsistent evaluations. Deep learning can enhance this process by automating using sophisticated algorithms assess both visual non-visual attributes, thereby increasing efficiency, accuracy, consistency. This study presents two standalone CNN-based methodologies automated quality grading, leveraging DenseNet121 InceptionV3 with custom layers tailored classification. A dataset 2,000 high-resolution images, manually curated from farms augmented diversity, was used training validation. Eight CNN architectures - DenseNet121, EfficientNetB4, InceptionResNetV2, InceptionV3, MobileNetV2, ResNet50, VGG16, VGG19 were evaluated. Experimental findings showed achieved the highest accuracy (95.67%) strong precision/recall scores (96%), making them most promising models. Meanwhile, MobileNetV2 identified as fastest model terms classification speed; however, its relatively low limits practical application tasks. while marginally slower at 0.015 0.011 seconds per image, respectively, offered a good balance between computational cost elevated accuracy. excels feature reuse through dense connectivity, reducing redundancy improving performance on smaller datasets, utilizes multi-scale extraction capture intricate patterns effectively. Both models demonstrate robustness under varying conditions, reliability deployment scenarios. highlights potential CNNs provide reliable, scalable solution benefiting farmers expanding opportunities.

Language: Английский

---

Recent innovations in machine learning for skin cancer lesion analysis and classification: A comprehensive analysis of computer‐aided diagnosis

Precision Medical Sciences, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 17, 2025

Abstract The global primary health concern of skin cancer emphasizes the need for quick and accurate diagnosis to improve patient outcomes. Although, it might be challenging evaluate possible risk a spot merely by looking at feeling it. This review article offers thorough overview current breakthroughs in machine learning (ML) computer‐aided diagnostics (CAD) aim analysis classification lesions over past 6 years. paper carefully reviews whole diagnostic process: data preparation, lesion segmentation, feature extraction, selection, final classification. Analyzed are many publicly accessible datasets creative ideas including deep (DL) ML integrated with computer vision, together their impact on increasing accuracy. Given variety complexity lesions, even enormous progress, there still major obstacles. rigorously assesses methods, notes areas great challenge, provides recommendations direct next research targeted improving early detection strategies CAD systems.

Language: Английский

---

Latest Frontiers of Machine, Deep, and Reinforcement Learning Algorithms for Cutting-Edge Applications

Published: Jan. 1, 2025

Language: Английский

---

Integrating Deep Learning and Energy Management Standards for Enhanced Solar–Hydrogen Systems: A Study Using MobileNetV2, InceptionV3, and ISO 50001:2018

Hydrogen, Journal Year: 2024, Volume and Issue: 5(4), P. 819 - 850

Published: Nov. 10, 2024

This study addresses the growing need for effective energy management solutions in university settings, with particular emphasis on solar–hydrogen systems. The study’s purpose is to explore integration of deep learning models, specifically MobileNetV2 and InceptionV3, enhancing fault detection capabilities AIoT-based environments, while also customizing ISO 50001:2018 standards align unique needs academic institutions. Our research employs comparative analysis two models terms their performance detecting solar panel defects assessing accuracy, loss values, computational efficiency. findings reveal that achieves 80% making it suitable resource-constrained InceptionV3 demonstrates superior accuracy 90% but requires more resources. concludes both offer distinct advantages based application scenarios, emphasizing importance balancing efficiency when selecting appropriate system management. highlights critical role continuous improvement leadership commitment successful implementation universities.

Language: Английский

---