Generating Synthetic Medical Images by Using GAN to Improve CNN Performance in Skin Cancer Classification

Sedigh, Pooyan; Sadeghian, Rasoul; Masouleh, Mehdi Tale

doi:10.1109/icrom48714.2019.9071823

Cited by 48 publications

(21 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…(Ech-Cherif, Misbahuddin, and Ech-Cherif 2019) in this research work CNN and DNN models are analysed deeplyfor skin cancer detection and the accuracy achieved by CNN model showed better performance compared to DNN model and analysis is done for RGB images of skin cancer using CNN model (Jayalakshmi and Sathiesh Kumar 2019). In this study it is proved that by comparing CNN and BN (Batch Normalization) CNN models with the results of accuracy, loss, precision, recall and F1 score values the BN CNN based system produces acceptable results and analysis is done for CT images of skin cancer detection (Sedigh, Sadeghian, and Masouleh 2019). in this study it is demonstrated by comparing CNN and GAN (Generative Adversarial Network) algorithms with the results of accuracy, sensitivity, specificity and F1 score values the CNN algorithm showing better performance compared to DNN algorithm.…”

mentioning

confidence: 85%

Detection of Skin Cancer using Inception V3 And Inception V4 Convolutional Neural Network (CNN) For Accuracy Improvement

Nandini¹,

Puviarasi²

2021

revistageintec

View full text Add to dashboard Cite

Aim:The main aim of this work is to measure the accuracy for automated detection of dermal cell images using the Convolutional Neural Network (CNN) algorithm. Materials and Methods:The skin images the dataset collected from International skin images collaboration (ISIC). In this framework 1200 images are used out of which (80%) are trained and (20%) are used for testing for the detection of skin cancer. 1200 images are used for group I (Inception V4) in comparison with Inception V3 and statistical analysis done using SPSS analysis. The sample size of two groups is calculated using G power with pretest power of 80 and alpha value 0.05 (error rate) with inputs 2400 (1200*2). Results: The inception V4 using CNN shows better results in mean Accuracy of 92.34±0.87 followed by inception V3 produces an accuracy of 90.34±0.13 with the significance value of <0.001. Conclusion: It is concluded that based on the execution analysis, the Inception V4 appears to be better accuracy compared with the Inception V3 algorithm.

show abstract

mentioning

confidence: 85%

Detection of Skin Cancer using Inception V3 And Inception V4 Convolutional Neural Network (CNN) For Accuracy Improvement

Nandini¹,

Puviarasi²

2021

revistageintec

View full text Add to dashboard Cite

show abstract

“…Besides, ∇ it is stochastic gradient descent employed in training GANs adhering to θd and θg parameters. About algorithm one and compensating the training data scarcity, Sedigh et al [81] put forth a CNN algorithm, a GAN variant towards generating mock images of skin cancer. Fig.…”

Section: Deep Learning and Generative Adversarial Network (Gan)mentioning

confidence: 99%

Skin Lesion Classification Based on Deep Convolutional Neural Networks Architectures

Saeed

Zeebaree

2021

JASTT

View full text Add to dashboard Cite

Skin cancer is among the primary cancer types that manifest due to various dermatological disorders, which may be further classified into several types based on morphological features, color, structure, and texture. The mortality rate of patients who have skin cancer is contingent on preliminary and rapid detection and diagnosis of malignant skin cancer cells. Limitations in current dermoscopic images, including shadow, artifact, and noise, affect image quality, which may hamper detection effort. Attempts to overcome these challenges have been made by analyzing the images using deep learning neural networks to perform skin cancer detection. In this paper, the authors review the state-of-the-art in authoritative deep learning concepts pertinent to skin cancer detection and classification.

show abstract

“…The variety of dragon motifs on Goutou makes it difficult for humans to quickly remember each one F I G U R E 3 Large defects can adversely affect the identification of the dragon motifs Deep learning can be used not only for defect detection but also to generate defect images for augmenting datasets. 27,28 The applied algorithms are Generating Adversary Networks (GANs). GAN was originally proposed by Goodfellow et al 29 GANs can be described as a powerful data distribution fitter, consisting mainly of a generator and a discriminator.…”

Section: Introductionmentioning

confidence: 99%

“…Deep learning can be used not only for defect detection but also to generate defect images for augmenting datasets 27,28 . The applied algorithms are Generating Adversary Networks (GANs).…”

Section: Introductionmentioning

confidence: 99%

Automatic segmentation, inpainting, and classification of defective patterns on ancient architecture using multiple deep learning algorithms

Zou

Zhao

2021

Struct Control Health Monit

View full text Add to dashboard Cite

Ancient architectures have a lot of distinctive patterns on their component surfaces. But due to the wind and the Sun, the patterns will present extensive deficiencies, which are not conducive to the routine inspections and the subsequent repair work. To overcome these limits, an automatic deep learning-based method of segmentation, inpainting, and classification for defective patterns on ancient architecture is proposed. First, You Only Look At CoefficienTs, which is a real-time instance segmentation network, is applied to obtain the mask of the defective parts. Then Generative Image Inpainting with Contextual Attention, which is an image inpainting algorithm, is used to reconstruct the defective parts. Finally, Residual Neural Network-50 is used to classify the reconstructed images. In this paper, three types of dragon motifs in the Forbidden City were studied. The results show that the classification accuracy of the reconstructed images is increased by an average of 13.1%, with the maximum increasing by 23.5%. The proposed methods can prepare for future routine inspections in advance.

show abstract

Generating Synthetic Medical Images by Using GAN to Improve CNN Performance in Skin Cancer Classification

Cited by 48 publications

References 28 publications

Detection of Skin Cancer using Inception V3 And Inception V4 Convolutional Neural Network (CNN) For Accuracy Improvement

Detection of Skin Cancer using Inception V3 And Inception V4 Convolutional Neural Network (CNN) For Accuracy Improvement

Skin Lesion Classification Based on Deep Convolutional Neural Networks Architectures

Automatic segmentation, inpainting, and classification of defective patterns on ancient architecture using multiple deep learning algorithms

Contact Info

Product

Resources

About