Development of Convolutional Neural Network-Based AI-Dermatoscope for Non-Invasive Skin Assessments DOI
Nipun Shantha Kahatapitiya, Akila Wijethunge, Sajith Edirisinghe

и другие.

Опубликована: Дек. 7, 2023

Early detection of skin conditions is crucial, and some can become more difficult to treat if left untreated. The gold standard Dermatoscope a non-invasive technique used for the examination evaluation lesions, which equipped with magnifying lens light source. However, precise inspection existing dermatoscopes has limitation due unavailability image-analyzing methods. Herein, this study reports successful development Convolutional Neural Networks (CNN) based, Artificial intelligence (AI)-Dermatoscope integrating optics smart illumination system enhance accurate acne skin. was trained on large dataset accurately identify classify conditions. Finally, utilizes CNN knowledge predict new images provide diagnostic information doctors other healthcare professionals. Thus, will improve accuracy speed diagnosis, consequently, health-related quality life patients.

Язык: Английский

A comprehensive review of model compression techniques in machine learning DOI Creative Commons
Pierre V. Dantas,

Waldir Sabino da Silva,

Lucas C. Cordeiro

и другие.

Applied Intelligence, Год журнала: 2024, Номер 54(22), С. 11804 - 11844

Опубликована: Сен. 2, 2024

Abstract This paper critically examines model compression techniques within the machine learning (ML) domain, emphasizing their role in enhancing efficiency for deployment resource-constrained environments, such as mobile devices, edge computing, and Internet of Things (IoT) systems. By systematically exploring lightweight design architectures, it is provided a comprehensive understanding operational contexts effectiveness. The synthesis these strategies reveals dynamic interplay between performance computational demand, highlighting balance required optimal application. As models grow increasingly complex data-intensive, demand resources memory has surged accordingly. escalation presents significant challenges artificial intelligence (AI) systems real-world applications, particularly where hardware capabilities are limited. Therefore, not merely advantageous but essential ensuring that can be utilized across various domains, maintaining high without prohibitive resource requirements. Furthermore, this review underscores importance sustainable development. introduction hybrid methods, which combine multiple techniques, promises to deliver superior efficiency. Additionally, development intelligent frameworks capable selecting most appropriate strategy based on specific application needs crucial advancing field. practical examples engineering applications discussed demonstrate impact techniques. optimizing complexity efficiency, ensures advancements AI technology remain widely applicable. thus contributes academic discourse guides innovative solutions efficient responsible practices, paving way future Graphical abstract

Язык: Английский

Процитировано

15
Skin Cancer Detection and Classification Using Neural Network Algorithms: A Systematic Review DOI Creative Commons
Pamela Hermosilla, Ricardo Soto, Emanuel Vega

и другие.

Diagnostics, Год журнала: 2024, Номер 14(4), С. 454 - 454

Опубликована: Фев. 19, 2024

In recent years, there has been growing interest in the use of computer-assisted technology for early detection skin cancer through analysis dermatoscopic images. However, accuracy illustrated behind state-of-the-art approaches depends on several factors, such as quality images and interpretation results by medical experts. This systematic review aims to critically assess efficacy challenges this research field order explain usability limitations highlight potential future lines work scientific clinical community. study, was carried out over 45 contemporary studies extracted from databases Web Science Scopus. Several computer vision techniques related image video processing diagnosis were identified. context, focus process included algorithms employed, result accuracy, validation metrics. Thus, yielded significant advancements using deep learning machine algorithms. Lastly, establishes a foundation research, highlighting contributions opportunities improve effectiveness learning.

Язык: Английский

Процитировано

8
Fusion of transformer attention and CNN features for skin cancer detection DOI
Hatice Çatal Reis, Veysel Turk

Applied Soft Computing, Год журнала: 2024, Номер 164, С. 112013 - 112013

Опубликована: Июль 18, 2024

Язык: Английский

Процитировано

6
Advanced deep learning and large language models: Comprehensive insights for cancer detection DOI
Yassine Habchi, Hamza Kheddar, Yassine Himeur

и другие.

Image and Vision Computing, Год журнала: 2025, Номер unknown, С. 105495 - 105495

Опубликована: Март 1, 2025

Язык: Английский

Процитировано

0
Preprocessed Vision Transformers and Classical Classifiers in Diagnosing Skin Diseases DOI Open Access
Bilal Şenol, Uğur Demiroğlu

DÜMF Mühendislik Dergisi, Год журнала: 2025, Номер 16(1), С. 69 - 80

Опубликована: Март 26, 2025

Vision Transformers (ViTs) are the state-of-the-art deep learning technology in medicine. ViTs require a large number of parameters, so they need relatively dataset for learning. This is currently possible due to digitization healthcare. As comparison, we also use classical classifiers, which characterized by low input data. In clinical practice, high-resolution images such as those from dermoscopy, confocal microscopy, reflectance and Raman spectroscopy used diagnose skin diseases. have potential practice. The advantage model over convolutional neural networks that do not operations. Preprocessed were classified experimentally using five models various sizes respective classifiers. Comparative experiments conducted on preprocessed dermatoscopic another dataset. article introduces an artificial intelligence method identifying conditions. contains into 5 categories: normal, melanoma, arsenic, psoriasis, eczema. During study, underwent initial processing Adaptive Histogram Equalization (AHE) technique, enhanced contrast reveal important details. Following this preprocessing, features obtained ViTs, renowned their ability capture intricate visual information. These extracted then utilized conjunction with traditional machine resulting accurate diagnosis conditions being studied. findings emphasize effectiveness combining classifiers tasks related medical image classification.

Язык: Английский

Процитировано

0
Automated explainable deep learning framework for multiclass skin cancer detection and classification using hybrid YOLOv8 and vision transformer (ViT) DOI
Humam AbuAlkebash, Radhwan A. A. Saleh, H. Metin Ertunç

и другие.

Biomedical Signal Processing and Control, Год журнала: 2025, Номер 108, С. 107934 - 107934

Опубликована: Апрель 29, 2025

Язык: Английский

Процитировано

0
DEEPSCAN: Integrating Vision Transformers for Advanced Skin Lesion Diagnostics DOI Open Access

A Jenefa,

Edward Naveen,

Vinayakumar Ravi

и другие.

The Open Dermatology Journal, Год журнала: 2024, Номер 18(1)

Опубликована: Март 21, 2024

Introduction/Background The rise in dermatological conditions, especially skin cancers, highlights the urgency for accurate diagnostics. Traditional imaging methods face challenges capturing complex lesion patterns, risking misdiagnoses. Classical CNNs, though effective, often miss intricate patterns and contextual nuances. Materials Methods Our research investigates adoption of Vision Transformers (ViTs) diagnosing lesions, capitalizing on their attention mechanisms global insights. Utilizing fictional Dermatological Dataset (DermVisD) with over 15,000 annotated images, we compare ViTs against traditional CNNs. This approach aims to assess potential benefits dermatology. Results Initial experiments showcase an 18% improvement diagnostic accuracy using achieving a remarkable 97.8% validation set. These findings suggest that are significantly more adept at recognizing patterns. Discussion integration into marks promising shift towards By leveraging understanding mechanisms, offer nuanced could surpass methods. advancement indicates setting new benchmarks Conclusion present significant field imaging, potentially redefining reliability standards. study underscores transformative impact detection diagnosis advocating broader clinical settings.

Язык: Английский

Процитировано

3
A novel Skin lesion prediction and classification technique: ViT‐GradCAM DOI Creative Commons
Muhammad Shafiq, Kapil Aggarwal,

J. Jayachandran

и другие.

Skin Research and Technology, Год журнала: 2024, Номер 30(9)

Опубликована: Сен. 1, 2024

Skin cancer is one of the highly occurring diseases in human life. Early detection and treatment are prime necessary points to reduce malignancy infections. Deep learning techniques supplementary tools assist clinical experts detecting localizing skin lesions. Vision transformers (ViT) based on image segmentation classification using multiple classes provide fairly accurate gaining more popularity due legitimate multiclass prediction capabilities.

Язык: Английский

Процитировано

3
Systematic Review of Deep Learning Techniques in Skin Cancer Detection DOI Creative Commons
Carolina Magalhaes, Joaquim Mendes, Ricardo Vardasca

и другие.

BioMedInformatics, Год журнала: 2024, Номер 4(4), С. 2251 - 2270

Опубликована: Ноя. 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.

Язык: Английский

Процитировано

2
Dermo-Seg: ResNet-UNet Architecture and Hybrid Loss Function for Detection of Differential Patterns to Diagnose Pigmented Skin Lesions DOI Creative Commons

Sannia Arshad,

Tehmina Amjad,

Ayyaz Hussain

и другие.

Diagnostics, Год журнала: 2023, Номер 13(18), С. 2924 - 2924

Опубликована: Сен. 12, 2023

Convolutional neural network (CNN) models have been extensively applied to skin lesions segmentation due their information discrimination capabilities. However, CNNs' struggle capture the connection between long-range contexts when extracting deep semantic features from lesion images, resulting in a gap that causes distortion lesions. Therefore, detecting presence of differential structures such as pigment networks, globules, streaks, negative and milia-like cysts becomes difficult. To resolve these issues, we proposed an approach based on semantic-based (Dermo-Seg) detect using UNet model with transfer-learning-based ResNet-50 architecture hybrid loss function. The Dermo-Seg uses backbone encoder model. We combination focal Tversky IOU functions handle dataset's highly imbalanced class ratio. obtained results prove intended performs well compared existing models. dataset was acquired various sources, ISIC18, ISBI17, HAM10000, evaluate dealt data imbalance present within each at pixel level our achieves mean score 0.53 for 0.67 0.66 0.58 milia-like-cysts. Overall, is efficient different achieved 96.4% index. Our system improves index most recent network.

Язык: Английский

Процитировано

6