Segmentation of Skin Lesions and their Attributes in Dermatoscopic Images Based on Convolutional Neural Networks

Antonio Vinicius de Moura Rodrigues; Roberta Barbosa Oliveira

doi:10.22456/2175-2745.143546

Authors

Antonio Vinicius de Moura Rodrigues University of Brasilia (UnB) https://orcid.org/0000-0002-0332-6642
Roberta Barbosa Oliveira University of Brasilia (UnB)

DOI:

https://doi.org/10.22456/2175-2745.143546

Keywords:

Deep Learning, Convolutional Neural Networks, TernausNet, MASK R-CNN, Image Segmentation, Skin Cancer

Abstract

Segmentation of skin lesions in dermoscopic images is a crucial step in diagnosing skin cancer, and Convolutional Neural Networks (CNNs) have emerged as powerful tools to address this challenge. This paper evaluated the effectiveness of two CNN models, TernausNet-16 and Mask R-CNN, in segmenting skin lesions and five of their attributes in the dermatoscopic images from the ISIC 2018 Challenge dataset. Jaccard Similarity Index (JSI) and Dice Similarity Coefficient (DSC) have been used as evaluation metrics. The results revealed that Mask R-CNN significantly outperformed TernausNet-16. The best model achieved 82.57% JSI and 84.76% DSC for lesion segmentation, and 42.86% JSI and 51.60% DSC for attribute segmentation when addressing attribute imbalance. Despite the longer training time, the results highlighted the potential of Mask R-CNN for improving the effectiveness of melanoma segmentation.

Downloads

Download data is not yet available.

References

IARC, I. A. for Research on C. Skin cancer. 2023. Accessed on: July 14, 2024. Available at: ⟨https://www.iarc.who.int/cancer-type/skin-cancer/#:∼:text=Skin%20cancers%20are%20the%20most,people%20died%20from%20the%20disease⟩.

MIRIKHARAJI, Z. et al. A survey on deep learning for skin lesion segmentation. Medical Image Analysis, Elsevier BV, v. 88, p. 102863, Aug. 2023. ISSN 1361-8415. Available at: ⟨http://dx.doi.org/10.1016/j.media.2023.102863⟩.

HESAMIAN, M. H. et al. Deep Learning Techniques for Medical Image Segmentation: Achievements and Challenges. Journal of Digital Imaging, v. 32, p. 582–596, 2019. ISSN 1618-727X. Available at: ⟨https://doi.org/10.1007/s10278-019-00227-x⟩.

JAHANIFAR, M. et al. Segmentation of Skin Lesions and their Attributes Using Multi-Scale Convolutional Neural Networks and Domain Specific Augmentations. CoRR, 2019. Available at: ⟨https://arxiv.org/abs/1809.10243⟩.

LÓPEZ-LABRACA, J. et al. An interpretable CNN-based CAD system for skin lesion diagnosis. Artificial Intelligence in Medicine, v. 132, p. 102370, 2022. ISSN 0933-3657. Available at: ⟨https://www.sciencedirect.com/science/article/pii/S0933365722001300⟩.

RONNEBERGER, O.; FISCHER, P.; BROX, T. U-Net: Convolutional Networks for Biomedical Image Segmentation. CoRR, abs/1505.04597, 2015. Available at: ⟨http://arxiv.org/abs/1505.04597⟩.

IGLOVIKOV, V.; SHVETS, A. TernausNet: U-Net with VGG11 Encoder Pre-Trained on ImageNet for Image Segmentation. CoRR, abs/1801.05746, 2018. Available at: ⟨http://arxiv.org/abs/1801.05746⟩.

SIMONYAN, K.; ZISSERMAN, A. Very Deep Convolutional Networks for Large-Scale Image Recognition. CoRR, abs/1409.1556, 2014. Available at: ⟨https://api.semanticscholar.org/CorpusID:14124313⟩.

SHVETS, A. A. et al. Automatic Instrument Segmentation in Robot-Assisted Surgery using Deep Learning. In: 2018 17th IEEE International Conference on Machine Learning and Applications (ICMLA). IEEE, 2018. Available at: ⟨http://dx.doi.org/10.1109/ICMLA.2018.00100⟩.

HE, K. et al. Mask R-CNN. In: 2017 IEEE International Conference on Computer Vision (ICCV). [S.l.: s.n.], 2017. p. 2980–2988.

REN, S. et al. Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. CoRR, abs/1506.01497, 2015. Available at: ⟨http://arxiv.org/abs/1506.01497⟩.

ARSHAD, S. et al. Dermo-Seg: ResNet-UNet Architecture and Hybrid Loss Function for Detection of Differential Patterns to Diagnose Pigmented Skin Lesions. Diagnostics, v. 13, n. 18, 2023. ISSN 2075-4418. Available at: ⟨https://www.mdpi.com/2075-4418/13/18/2924⟩.

ALFARO, E. et al. A Brief Analysis of U-Net and Mask R-CNN for Skin Lesion Segmentation. In: . [S.l.: s.n.], 2019. p. 000123–000126.

QIAN, C. et al. A Detection and Segmentation Architecture for Skin Lesion Segmentation on Dermoscopy Images. CoRR, abs/1809.03917, 2018. Available at: ⟨http://arxiv.org/abs/1809.03917⟩.

CHEN, E. Z. et al. Lesion Attributes Segmentation for Melanoma Detection with Multi-Task U-Net. In: 2019 IEEE 16th International Symposium on Biomedical Imaging (ISBI 2019). [S.l.: s.n.], 2019. p. 485–488.

TSCHANDL, P.; ROSENDAHL, C.; KITTLER, H. The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions. Scientific Data, v. 5, n. 1, p. 180161, 2018. ISSN 2052-4463. Available at: ⟨https://doi.org/10.1038/sdata.2018.161⟩.

CODELLA, N. C. F. et al. Skin Lesion Analysis Toward Melanoma Detection 2018: A Challenge Hosted by the International Skin Imaging Collaboration (ISIC). CoRR, abs/1902.03368, 2019. Available at: ⟨http://arxiv.org/abs/1902.03368⟩.

PARSANIA, P.; VIRPARIA, P. A review: Image interpolation techniques for image scaling. International Journal of Innovative Research in Computer and Communication Engineering, v. 02, p. 7409–7414, 01 2015.

DENG, J. et al. ImageNet: A large-scale hierarchical image database. p. 248–255, June 2009. ISSN 1063-6919.

LI, L.; SEO, W. Deep Learning and Transfer Learning for Skin Cancer Segmentation and Classification. In: 2021 IEEE 21st International Conference on Bioinformatics and Bioengineering (BIBE). [S.l.: s.n.], 2021. p. 1–5.

LIN, T. et al. Microsoft COCO: Common Objects in Context. CoRR, abs/1405.0312, 2014. Available at: ⟨http://arxiv.org/abs/1405.0312⟩.

HE, K. et al. Deep Residual Learning for Image Recognition. CoRR, abs/1512.03385, 2015. Available at: ⟨http://arxiv.org/abs/1512.03385⟩.

PEREZ, F. et al. Data Augmentation for Skin Lesion Analysis. In: Lecture Notes in Computer Science. Springer International Publishing, 2018. p. 303–311. Available at: ⟨https://doi.org/10.1007%2F978-3-030-01201-4_33⟩.

ALOMAR, K.; AYSEL, H.; CAI, X. Data augmentation in classification and segmentation: A survey and new strategies. Journal of Imaging, v. 9, p. 46, 02 2023.

OKTAY, O. et al. Attention U-Net: Learning Where to Look for the Pancreas. 2018. Available at: ⟨https://arxiv.org/abs/1804.03999⟩.