Skin Cancer Detection from Low-Resolution Images Using Transfer Learning

Khan, Mudasir; Uddin, Abdul Hasib; Nahid, Abdullah-Al; Bairagi, Anupam Kumar

doi:10.1007/978-981-16-2422-3_26

Cited by 5 publications

(3 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The accuracy is different in all studies, as different datasets (HAM10000, Kaggle and clinical images) are used. With HAM10000 dataset, Khan et al [ 8 ] achieved an accuracy of 80.46% using VGG16 model architecture, and Agrahari et al [ 10 ] and Chaturvedi et al [ 11 ] had achieved accuracy rates of 80.81% and 83.10%, respectively on MobileNet model architecture.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

An Enhanced Transfer Learning Based Classification for Diagnosis of Skin Cancer

et al. 2022

View full text Add to dashboard Cite

Skin cancer is the most commonly diagnosed and reported malignancy worldwide. To reduce the death rate from cancer, it is essential to diagnose skin cancer at a benign stage as soon as possible. To save lives, an automated system that can detect skin cancer in its earliest stages is necessary. For the diagnosis of skin cancer, various researchers have performed tasks using deep learning and transfer learning models. However, the existing literature is limited in terms of its accuracy and its troublesome and time-consuming process. As a result, it is critical to design an automatic system that can deliver a fast judgment and considerably reduce mistakes in diagnosis. In this work, a deep learning-based model has been designed for the identification of skin cancer at benign and malignant stages using the concept of transfer learning approach. For this, a pre-trained VGG16 model is improved by adding one flatten layer, two dense layers with activation function (LeakyReLU) and another dense layer with activation function (sigmoid) to enhance the accuracy of this model. This proposed model is evaluated on a dataset obtained from Kaggle. The techniques of data augmentation are applied in order to enhance the random-ness among the input dataset for model stability. The proposed model has been validated by considering several useful hyper parameters such as different batch sizes of 8, 16, 32, 64, and 128; different epochs and optimizers. The proposed model is working best with an overall accuracy of 89.09% on 128 batch size with the Adam optimizer and 10 epochs and outperforms state-of-the-art techniques. This model will help dermatologists in the early diagnosis of skin cancers.

show abstract

Section: Resultsmentioning

confidence: 99%

“…They obtained an accuracy of 96.0%. Khan et al [ 8 ] used various pre-trained architectures like VGG16, DenseNet169, DenseNet161, and ResNet50. In this, they pushed the boundary of neural networks by using low resolution pixels such as 80 × 80, 64 × 64, and 32 × 32.…”

Section: Introductionmentioning

confidence: 99%

An Enhanced Transfer Learning Based Classification for Diagnosis of Skin Cancer

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Transfer learning using DenseNet has been explored for skin cancer detection. Khan et al utilized a low-resolution, highly imbalanced, grayscale HAM10000 skin cancer dataset and applied transfer learning with DenseNet169, achieving a performance of 78.56% for 64 × 64 pixel images [21]. Panthakkan et al developed a unique deep-learning model that integrates Xception and ResNet50, achieving a prediction accuracy of 97.8% [22].…”

Section: Transfer Learning In Skin Cancer Detection and Classificationmentioning

confidence: 99%

Enhancing Skin Cancer Detection and Classification: Exploring the Impact of Attention Mechanisms in Transfer Learning Models

Halabi

2023

Preprint

View full text Add to dashboard Cite

Skin cancer is a global health concern with increasing prevalence, necessitating effective early detection and classification systems. Transfer learning has emerged as a powerful tool in this domain, enhancing diagnostic accuracy. Attention mechanisms, which selectively focus on pertinent image features, play a pivotal role in transfer learning. While previous studies have highlighted their effectiveness, their applicability across different models and datasets remains understudied. This empirical study explores the impact of various attention mechanisms on the performance of transfer learning models for skin cancer detection and classification. Using five transfer learning models (DenseNet121, InceptionV3, MobileNet, VGG16, and Xception) and six attention mechanisms (Channel Attention, Global Context Attention, Guided Attention, Nonlocal Attention, Positional Attention, and Spatial Attention), we conducted 105 experiments across three datasets. Traditional metrics (accuracy, precision, recall, and f1-score) were employed for empirical validation. The results reveal a nuanced relationship between attention mechanisms, transfer learning models, and datasets. Overall, attention mechanisms exhibit the potential to enhance skin cancer classification. Spatial and channel attention mechanisms consistently outperform others, offering simplicity and effectiveness. Model-specific selection of attention mechanisms is crucial, with a trade-off between model complexity and performance evident. This study provides insights into developing efficient skin cancer classification models utilizing attention mechanisms.

show abstract