Diagnosis of Various Skin Cancer Lesions Based on Fine-Tuned ResNet50 Deep Network
Source abstract
Abstract
With the massive success of deep networks, there have been significant efforts to analyze cancer diseases, especially skin cancer. For this purpose, this work investigates the capability of deep networks in diagnosing a variety of dermoscopic lesion images. This paper aims to develop and fine-tune a deep learning architecture to diagnose different skin cancer grades based on dermatoscopic images. Fine-tuning is a powerful method to obtain enhanced classification results by the customized pre-trained network. Regularization, batch normalization, and hyperparameter optimization are performed for fine-tuning the proposed deep network. The proposed fine-tuned ResNet50 model successfully classified 7-respective classes of dermoscopic lesions using the publicly available HAM10000 dataset. The developed deep model was compared against two powerful models, i.e., InceptionV3 and VGG16, using the Dice similarity coefficient (DSC) and the area under the curve (AUC). The evaluation results show that the proposed model achieved higher results than some recent and robust models.
Source: openalex_inverted_index · Quality: baseline
Research context
OpenAlex topics: Cutaneous Melanoma Detection and Management · Nonmelanoma Skin Cancer Studies · AI in cancer detection
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
References (45)
Show source-provided references
1. Comparing artificial intelligence algorithms to 157 German dermatologists: The melanoma classification benchmark10.1016/j.ejca.2018.12.016 ↗
2. Computer-aided diagnosis of skin diseases using deep neural networks10.3390/app10072488 ↗
3. A convolutional neural network trained with dermoscopic images performed on par with 145 dermatologists in a clinical melanoma image classification task10.1016/j.ejca.2019.02.005 ↗
4. Deep learning-based system for automatic melanoma detection10.1109/ACCESS.2019.2962812 ↗
5. Deep learning outperformed 11 pathologists in the classification of histopathological melanoma images10.1016/j.ejca.2019.06.012 ↗
6. Step-wise integration of deep class-specific learning for dermoscopic image segmentation10.1016/j.patcog.2018.08.001 ↗
7. Domain-specific classification-pretrained fully convolutional network encoders for skin lesion segmentation10.1016/j.compbiomed.2018.11.010 ↗
8. Skin cancer detection based on deep learning and entropy to detect outlier samples
9. Fusing fine-tuned deep features for skin lesion classification10.1016/j.compmedimag.2018.10.007 ↗
10. Skin lesion classification using hybrid deep neural networks
11. Adaptive melanoma diagnosis using evolving clustering, ensemble and deep neural networks10.1016/j.knosys.2019.06.015 ↗
12. Improvements to deep convolutional neural networks for LVCSR
13. Delving deep into rectifiers: Surpassing human-level performance on imagenet classification
14. Improving deep neural networks using softplus units
15. Learning to learn without gradient descent by gradient descent
16. Practical recommendations for gradient-based training of deep architectures
17. Generalization error in deep learning
18. Greedy layer-wise training of deep networks
19. Dermoscopy diagnosis of cancerous lesions utilizing dual deep learning algorithms via visual and audio (sonification) outputs: Laboratory and prospective observational studies10.1016/j.ebiom.2019.01.028 ↗
20. Histology of non-melanoma skin cancers: An update10.3390/biomedicines5040071 ↗
21. Deep neural networks are superior to dermatologists in melanoma image classification10.1016/j.ejca.2019.05.023 ↗
22. Enhanced classifier training to improve precision of a convolutional neural network to identify images of skin lesions10.1371/journal.pone.0218713 ↗
23. Skin lesions classification into eight classes for ISIC, 2019 using deep convolutional neural network and transfer learning10.1109/ACCESS.2020.3003890 ↗
24. Malignant melanoma classification using deep learning: Datasets, performance measurements, challenges and opportunities10.1109/ACCESS.2020.3001507 ↗
25. P. Tschandl, “The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions,” 2018. https://github.com/ptschandl/HAM10000_dataset (Accessed 3 February 2021).10.1038/sdata.2018.161 ↗
26. Competitive residual neural network for image classification10.1016/j.icte.2019.06.001 ↗
27. Deep residual learning for image recognition
28. S. R. Xiangyu Zhang Kaiming He and J. Sun, “Deep residual learning for image recognition,” 2015. https://arxiv.org/abs/1512.03385.
29. Detecting affect states using VGG16, ResNet50 and SE-ResNet50 networks10.1007/s42979-020-0114-9 ↗
30. Zouggar · Optimization in Machine Learning and Applications · 2020 · 3110.1007/978-981-15-0994-0_3 ↗
31. Optimization methods for large-scale machine learning10.1137/16M1080173 ↗
32. A comparative analysis of gradient descent-based optimization algorithms on convolutional neural networks
33. A survey of optimization methods from a machine learning perspective10.1109/TCYB.2019.2950779 ↗
34. Deep learning hyper-parameter optimization for video analytics in clouds10.1109/TSMC.2018.2840341 ↗
35. Skin lesion classification using hybrid deep neural networks
36. Improving skin-disease classification based on customized loss function combined with balanced mini-batch logic and real-time image augmentation10.1109/ACCESS.2020.3016653 ↗
37. A deep normalization and convolutional neural network for image smoke detection10.1109/ACCESS.2017.2747399 ↗
38. Influence of batch normalization on convolutional neural networks in HRRP target recognition
39. Analysis of deep networks with residual blocks and different activation functions: Classification of skin diseases
40. Going deep in medical image analysis: Concepts, methods, challenges and future directions10.1109/ACCESS.2019.2929365 ↗
41. Improved neural network using Integral-Relu based prevention activation for face detection
42. Convolutional neural networks: An overview and application in radiology10.1007/s13244-018-0639-9 ↗
43. Detection of melanoma using deep learning techniques
44. Optimizing deep CNN architectures for face liveness detection10.3390/e21040423 ↗
45. Automated segmentation and classification of hepatocellular carcinoma using fuzzy C-Means and SVM10.1007/978-3-319-33793-7_9 ↗