A hybrid CNN architecture for skin lesion classification using deep learning

Jasil, S. P. Godlin; Ulagamuthalvi, V.

doi:10.1007/s00500-023-08035-w

Cited by 10 publications

(11 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hence, to provide an equitable evaluation of the network's efficacy across various datasets, it is essential that the prediction task remains constant, hence enabling a comparative examination of the influence of input data on the accuracy of the model. Finally, framing the problem as a benign-vs-malignant classification task allows a fair comparison with several reputable methods addressing the skin lesion classification [12,17,18] that also addressed the problem using binary classification methods. Nevertheless, extending the proposed network for multiclass classification is trivial and can be achieved by modifying the output layers as needed.…”

Section: ) Classification Layermentioning

confidence: 99%

“…Various researchers have proposed new techniques for integrating metadata into their learning mechanisms and have continued to improve upon each other's work over time. In this study, we compared our proposed technique to three state-of-the-art methods, Jasil (2023) [18] + DMF, Ningrum (2021) [17], and Gessert (2020) [12], which also fuse metadata into their network architectures for skin lesion classification. The image encoding modules for these three baselines were reproduced as reported in their publications.…”

Section: E Network Comparisonmentioning

confidence: 99%

“…The image encoding modules for these three baselines were reproduced as reported in their publications. Note that the skin lesion classification method proposed by Jasil and Ulagamuthalvi [18] does not inherently integrate image metadata into the network, making it difficult to compare with other metadatafusing networks. Therefore, their proposed network was used as the image-encoder component in our proposed DMF network, hence referred to as Jasil (2023) + DMF.…”

Section: E Network Comparisonmentioning

confidence: 99%

“…The results of our comparison study are presented in Table 7. By comparing the performance of the proposed Deep-MetaForge (BEiT + DMF) network and three other published approaches (Jasil (2023) [18], Ningrum (2021) [17], and Gessert (2020) [12]), across all datasets, except for PH2, our DeepMetaForge network achieved the highest macro-average F1 score. As discussed earlier, the relevance and quality of metadata can have a significant impact on classification performance.…”

Section: E Network Comparisonmentioning

confidence: 99%

“…• We empirically evaluate our proposed framework on four public skin lesion image datasets and compare the classification performance with stand-alone backbones and state-of-the-art metadata-fusing methods for skin lesion classification [12,17,18]. • We empirically investigate the scalability of the proposed framework by analyzing the trade-off between the efficacy and efficiency of the models.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

DeepMetaForge: A Deep Vision-Transformer Metadata-Fusion Network for Automatic Skin Lesion Classification

Vachmanus,

Noraset,

Piyanonpong

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Skin cancer is a dangerous form of cancer that develops slowly in skin cells. Delays in diagnosing and treating these malignant skin conditions may have serious repercussions. Likewise, early skin cancer detection has been shown to improve treatment outcomes. This paper proposes DeepMetaForge, a deeplearning framework for skin cancer detection from metadata-accompanied images. The proposed framework utilizes BEiT, a vision transformer pre-trained as a masked image modeling task, as the image-encoding backbone. We further propose merging the encoded metadata with the derived visual characteristics while compressing the aggregate information simultaneously, simulating how photos with metadata are interpreted. The experiment results on four public datasets of dermoscopic and smartphone skin lesion images reveal that the best configuration of our proposed framework yields 87.1% macro-average F1 on average. The empirical scalability analysis further shows that the proposed framework can be implemented in a variety of machine-learning paradigms, including applications on low-resource devices and as services. The findings shed light on not only the possibility of implementing telemedicine solutions for skin cancer on a nationwide scale that could benefit those in need of quality healthcare but also open doors to many intelligent applications in medicine where images and metadata are collected together, such as disease detection from CT-scan images and patients' expression recognition from facial images.

show abstract

Section: ) Classification Layermentioning

confidence: 99%

Section: E Network Comparisonmentioning

confidence: 99%

Section: E Network Comparisonmentioning

confidence: 99%

Section: E Network Comparisonmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

DeepMetaForge: A Deep Vision-Transformer Metadata-Fusion Network for Automatic Skin Lesion Classification

Vachmanus,

Noraset,

Piyanonpong

et al. 2023

IEEE Access

View full text Add to dashboard Cite

show abstract

Skin lesion classification using transfer learning

Nivedhitha,

Kalpana,

Sidthik

et al. 2024

Soft Comput

View full text Add to dashboard Cite

An effective multiclass skin cancer classification approach based on deep convolutional neural network

Houssein,

Abdelkareem,

et al. 2024

Cluster Comput

View full text Add to dashboard Cite

Skin cancer is one of the most dangerous types of cancer due to its immediate appearance and the possibility of rapid spread. It arises from uncontrollably growing cells, rapidly dividing cells in one area of the body, invading other bodily tissues, and spreading throughout the body. Early detection helps prevent cancer progress from reaching critical levels, reducing the risk of complications and the need for more aggressive treatment options. Convolutional neural networks (CNNs) revolutionize skin cancer diagnosis by extracting intricate features from images, enabling an accurate classification of lesions. Their role extends to early detection, providing a powerful tool for dermatologists to identify abnormalities in their nascent stages, ultimately improving patient outcomes. This study proposes a novel deep convolutional neural network (DCNN) approach to classifying skin cancer lesions. The proposed DCNN model is evaluated using two unbalanced datasets, namely HAM10000 and ISIC-2019. The DCNN model is compared with other transfer learning models, including VGG16, VGG19, DenseNet121, DenseNet201, and MobileNetV2. Its performance is assessed using four widely used evaluation metrics: accuracy, recall, precision, F1-score, specificity, and AUC. The experimental results demonstrate that the proposed DCNN model outperforms other deep learning (DL) models that utilized these datasets. The proposed DCNN model achieved the highest accuracy with the HAM10000 and ISIC-2019 datasets, reaching $$98.5\%$$ 98.5 % and $$97.1\%$$ 97.1 % , respectively. These experimental results show how competitive and successful the DCNN model is in overcoming the problems caused by class imbalance and raising skin cancer classification accuracy. Furthermore, the proposed model demonstrates superior performance, particularly excelling in terms of accuracy, compared to other recent studies that utilize the same datasets, which highlights the robustness and effectiveness of the proposed DCNN.

show abstract

A hybrid CNN architecture for skin lesion classification using deep learning

Cited by 10 publications

References 37 publications

DeepMetaForge: A Deep Vision-Transformer Metadata-Fusion Network for Automatic Skin Lesion Classification

DeepMetaForge: A Deep Vision-Transformer Metadata-Fusion Network for Automatic Skin Lesion Classification

Skin lesion classification using transfer learning

An effective multiclass skin cancer classification approach based on deep convolutional neural network

Contact Info

Product

Resources

About