Detection of Skin Cancer Based on Skin Lesion Images Using Deep Learning

Healthcare, Journal Year: 2022, Volume and Issue: 10(7), P. 1183 - 1183

Published: June 24, 2022

An increasing number of genetic and metabolic anomalies have been determined to lead cancer, generally fatal. Cancerous cells may spread any body part, where they can be life-threatening. Skin cancer is one the most common types its frequency worldwide. The main subtypes skin are squamous basal cell carcinomas, melanoma, which clinically aggressive responsible for deaths. Therefore, screening necessary. One best methods accurately swiftly identify using deep learning (DL). In this research, method convolution neural network (CNN) was used detect two primary tumors, malignant benign, ISIC2018 dataset. This dataset comprises 3533 lesions, including malignant, nonmelanocytic, melanocytic tumors. Using ESRGAN, photos were first retouched improved. augmented, normalized, resized during preprocessing step. lesion could classified a CNN based on an aggregate results obtained after many repetitions. Then, multiple transfer models, such as Resnet50, InceptionV3, Inception Resnet, fine-tuning. addition experimenting with several models (the designed CNN, Resnet), study's innovation contribution use ESRGAN Our model showed comparable pretrained model. Simulations ISIC 2018 that suggested strategy successful. 83.2% accuracy rate achieved by in comparison Resnet50 (83.7%), InceptionV3 (85.8%), Resnet (84%) models.

Language: Английский

---

An Efficient Deep Learning-Based Skin Cancer Classifier for an Imbalanced Dataset

Diagnostics, Journal Year: 2022, Volume and Issue: 12(9), P. 2115 - 2115

Published: Aug. 31, 2022

Efficient skin cancer detection using images is a challenging task in the healthcare domain. In today's medical practices, time-consuming procedure that may lead to patient's death later stages. The diagnosis of at an earlier stage crucial for success rate complete cure. efficient task. Therefore, numbers skilful dermatologists around globe are not enough deal with healthcare. huge difference between data from various sector classes leads imbalance problems. Due issues, deep learning models often trained on one class more than others. This study proposes novel learning-based detector imbalanced dataset. Data augmentation was used balance overcome imbalance. Skin Cancer MNIST: HAM10000 dataset employed, which consists seven lesions. Deep widely disease through images. (AlexNet, InceptionV3, and RegNetY-320) were employed classify cancer. proposed framework also tuned combinations hyperparameters. results show RegNetY-320 outperformed InceptionV3 AlexNet terms accuracy, F1-score, receiver operating characteristic (ROC) curve both balanced datasets. performance better conventional methods. ROC value obtained 91%, 88.1%, 0.95, significantly those state-of-the-art method, achieved 85%, 69.3%, 0.90, respectively. Our assist identification, could save lives, reduce unnecessary biopsies, costs patients, dermatologists, professionals.

Language: Английский

---

A deep learning outline aimed at prompt skin cancer detection utilizing gated recurrent unit networks and improved orca predation algorithm

Biomedical Signal Processing and Control, Journal Year: 2023, Volume and Issue: 90, P. 105858 - 105858

Published: Dec. 22, 2023

Language: Английский

---

Skin Cancer Detection Using Deep Learning—A Review

Diagnostics, Journal Year: 2023, Volume and Issue: 13(11), P. 1911 - 1911

Published: May 30, 2023

Skin cancer is one the most dangerous types of and primary causes death worldwide. The number deaths can be reduced if skin diagnosed early. mostly using visual inspection, which less accurate. Deep-learning-based methods have been proposed to assist dermatologists in early accurate diagnosis cancers. This survey reviewed recent research articles on classification deep learning methods. We also provided an overview common deep-learning models datasets used for classification.

Language: Английский

---

Skin Lesion Analysis and Cancer Detection Based on Machine/Deep Learning Techniques: A Comprehensive Survey

Life, Journal Year: 2023, Volume and Issue: 13(1), P. 146 - 146

Published: Jan. 4, 2023

The skin is the human body’s largest organ and its cancer considered among most dangerous kinds of cancer. Various pathological variations in body can cause abnormal cell growth due to genetic disorders. These changes cells are very dangerous. Skin slowly develops over further parts because high mortality rate cancer, early diagnosis essential. visual checkup manual examination lesions tricky for determination Considering these concerns, numerous recognition approaches have been proposed With fast progression computer-aided systems, a variety deep learning, machine computer vision were merged medical samples uncommon lesion samples. This research provides an extensive literature review methodologies, techniques, applied date. survey includes preprocessing, segmentation, feature extraction, selection, classification recognition. results impressive but still, some challenges occur analysis complex rare features. Hence, main objective examine existing techniques utilized discovery by finding obstacle that helps researchers contribute future research.

Language: Английский

---

Melanoma Detection Using Deep Learning-Based Classifications

Healthcare, Journal Year: 2022, Volume and Issue: 10(12), P. 2481 - 2481

Published: Dec. 8, 2022

One of the most prevalent cancers worldwide is skin cancer, and it becoming more common as population ages. As a general rule, earlier cancer can be diagnosed, better. result success deep learning (DL) algorithms in other industries, there has been substantial increase automated diagnosis systems healthcare. This work proposes DL method for extracting lesion zone with precision. First, image enhanced using Enhanced Super-Resolution Generative Adversarial Networks (ESRGAN) to improve image's quality. Then, segmentation used segment Regions Interest (ROI) from full image. We employed data augmentation rectify disparity. The then analyzed convolutional neural network (CNN) modified version Resnet-50 classify lesions. analysis utilized an unequal sample seven kinds HAM10000 dataset. With accuracy 0.86, precision 0.84, recall F-score proposed CNN-based Model outperformed study's results by significant margin. study culminates improved diagnosing that benefits medical professionals patients.

Language: Английский

---

Melanoma segmentation: A framework of improved DenseNet77 and UNET convolutional neural network

International Journal of Imaging Systems and Technology, Journal Year: 2022, Volume and Issue: 32(6), P. 2137 - 2153

Published: May 17, 2022

Abstract Melanoma is the most fatal type of skin cancer which can cause death victims at advanced stage. Extensive work has been presented by researcher on computer vision for lesion localization. However, correct and effective melanoma segmentation still a tough job because extensive variations found in shape, color, sizes moles. Moreover, presence light brightness further complicates task. We have improved deep learning (DL)‐based approach, namely, DenseNet77‐based UNET model. More clearly, we introduced DenseNet77 network encoder unit approach to computing more representative set image features. The calculated keypoints are later segmented decoder used two standard datasets, ISIC‐2017 ISIC‐2018 evaluate performance proposed acquired accuracies 99.21% 99.51% respectively. confirmed through both quantitative qualitative results that robust lesions accurately recognize moles varying colors sizes.

Language: Английский

---

A novel hybrid Extreme Learning Machine and Teaching–Learning-Based​ Optimization algorithm for skin cancer detection

Healthcare Analytics, Journal Year: 2023, Volume and Issue: 3, P. 100161 - 100161

Published: March 16, 2023

Skin cancers, such as melanoma, can be difficult to spot in their early stages because they often resemble benign moles. Early detection of melanoma is crucial it increases the chances successful treatment and prevents cancer from spreading other areas body. Machine learning algorithms computer vision techniques are versatile for detecting melanoma. However, current research has limitations, inaccurate longer computation times. This paper proposes a novel hybrid Extreme Learning (ELM) Teaching–Learning-Based Optimization (TLBO) algorithm technique ELM single-hidden layer feed-forward neural network that trained quickly accurately, while TLBO an optimization used fine-tune network's parameters improved performance. Together, these classify skin lesions or malignant images, potentially improving accuracy.

Language: Английский

---

SkinNet-INIO: Multiclass Skin Lesion Localization and Classification Using Fusion-Assisted Deep Neural Networks and Improved Nature-Inspired Optimization Algorithm

Diagnostics, Journal Year: 2023, Volume and Issue: 13(18), P. 2869 - 2869

Published: Sept. 6, 2023

Background: Using artificial intelligence (AI) with the concept of a deep learning-based automated computer-aided diagnosis (CAD) system has shown improved performance for skin lesion classification. Although convolutional neural networks (DCNNs) have significantly many image classification tasks, it is still difficult to accurately classify lesions because lack training data, inter-class similarity, intra-class variation, and inability concentrate on semantically significant parts. Innovations: To address these issues, we proposed an learning best feature selection framework multiclass in dermoscopy images. The performs preprocessing step at initial contrast enhancement using new technique that based dark channel haze top–bottom filtering. Three pre-trained models are fine-tuned next trained transfer concept. In fine-tuning process, added removed few additional layers lessen parameters later selected hyperparameters genetic algorithm (GA) instead manual assignment. purpose hyperparameter GA improve performance. After that, deeper layer each network features extracted. extracted fused novel serial correlation-based approach. This reduces vector length serial-based approach, but there little redundant information. We anti-Lion optimization this issue. finally classified machine algorithms. Main Results: experimental process was conducted two publicly available datasets, ISIC2018 ISIC2019. Employing obtained accuracy 96.1 99.9%, respectively. Comparison also state-of-the-art techniques shows accuracy. Conclusions: successfully enhances cancer region. Moreover, framework. fusion version maintains shorten computational time.

Language: Английский

---

A survey on cancer detection via convolutional neural networks: Current challenges and future directions

Neural Networks, Journal Year: 2023, Volume and Issue: 169, P. 637 - 659

Published: Nov. 7, 2023

Language: Английский

---