AIRCC's International Journal of Computer Science and Information Technology

Open Access Open Access  Restricted Access Subscription Access

Self-Learning AI Framework for Skin Lesion Image Segmentation and Classification


Affiliations
1 Department of Cyber Physical Systems, Central Electronics Engineering Research Institute, CSIR Madras Complex, Taramani, Chennai, India
2 Dept. of Computer Science, University of Bridgeport, 126 Park Ave, Bridgeport, United States
 

Image segmentation and classification are the two main fundamental steps in pattern recognition. To perform medical image segmentation or classification with deep learning models, it requires training on large image dataset with annotation. The dermoscopy images (ISIC archive) considered for this work does not have ground truth information for lesion segmentation. Performing manual labelling on this dataset is time-consuming. To overcome this issue, self-learning annotation scheme was proposed in the two-stage deep learning algorithm. The two-stage deep learning algorithm consists of U-Net segmentation model with the annotation scheme and CNN classifier model. The annotation scheme uses a K-means clustering algorithm along with merging conditions to achieve initial labelling information for training the U-Net model. The classifier models namely ResNet-50 and LeNet-5 were trained and tested on the image dataset without segmentation for comparison and with the U-Net segmentation for implementing the proposed self-learning Artificial Intelligence (AI) framework. The classification results of the proposed AI framework achieved training accuracy of 93.8% and testing accuracy of 82.42% when compared with the two classifier models directly trained on the input images.

Keywords

Self-Learning Annotation Scheme, K-Means Clustering, U-Net, Deep Learning & Skin Lesion Image.
User
Notifications
Font Size

  • Zhang, Ling et al., (2018) “Self-learning to detect and segment cysts in lung CT images without manual annotation”, IEEE 15th International Symposium on Biomedical Imaging (ISBI), IEEE Xplore DOI: 10.1109/ISBI.2018.8363763, pp. 1100-1103.

  • S. Roy, A. Panda and R. Naskar, (2019) "Unsupervised Ground Truth Generation for Automated Brain EM Image Segmentation", 6th International Conference on Signal Processing and Integrated Networks (SPIN), DOI: 10.1109/SPIN.2019.8711724, pp.66-71.

  • Lin, Bill S., Kevin Michael, Shivam Kalra and Hamid R. Tizhoosh.,(2017) “Skin lesion segmentation: U-Nets versus clustering”, IEEE Symposium Series on Computational Intelligence (SSCI), DOI: 10.1109/SSCI.2017.8280804, pp. 1-7.

  • S. M. Jaisakthi, P. Mirunalini and C. Aravindan, (2017) “Automatic Skin Lesion Segmentation using Semi-supervised Learning Technique.” Online Ref: arXiv:1703.04301.

  • S. M. Jaisakthi, P. Mirunalini and C. Aravindan, (2018) "Automated skin lesion segmentation of dermoscopic images using GrabCut and k-means algorithms," IET Computer Vision, Vol. 12, No. 8, DOI: 10.1049/iet-cvi.2018.5289, pp. 1088-1095.

  • Youssef Filali, Sabri Abdelouahed, Abdellah Aarab, (2019) “An Improved Segmentation Approach for Skin Lesion Classification”, International journal of statistics, optimization and information computing, DOI: 10.19139/soic.v7i2.533, Vol 7, No 2, pp.456-467.

  • Manu Goyal et. al, (2019) “Skin lesion segmentation in dermoscopic images with ensemble deep learning methods”, Online Ref: arXiv:1902.00809v2 [eess.IV].

  • Kermany, Daniel S. et al., (2018) “Identifying Medical Diagnoses and Treatable Diseases by Image-Based Deep Learning.” Cell Vol. 172, No. 5, Elsevier, DOI: 10.1016/j.cell.2018.02.010, pp. 1122-1131.e9.

  • Lee, Cecilia S. et al., (2016) “Deep learning is effective for the classification of OCT images of normal versus Age-related Macular Degeneration.” Ophthalmology Retina, Vol. 1, No.4, DOI:10.1016/j.oret.2016.12.009, pp.322-327.

  • Ronneberger O., Fischer P., Brox T., (2015) “U-Net: Convolutional Networks for Biomedical Image Segmentation”, In: Navab N. et al., (Eds.): Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015. Lecture Notes in Computer Science, vol 9351. Springer, DOI: 10.1007/978-3-319-24574-4_28, pp.234-241.

  • A. Romero Lopez, X. Giro-i-Nieto, J. Burdick and O. Marques, (2017) "Skin lesion classification from dermoscopic images using deep learning techniques," 13th International Conference on Biomedical Engineering (BioMed), DOI: 10.2316/P.2017.852-053, pp. 49-54.

  • L. Yu, H. Chen, Q. Dou, J. Qin, and P.-A. Heng, (2017) “Automated melanoma recognition in dermoscopy images via very deep residual networks”, IEEE transactions on medical imaging, Vol. 36, No. 4, DOI: 10.1109/TMI.2016.2642839, pp. 994–1004.

  • Noel Codella, Veronica Rotemberg, Philipp Tschandl, M. Emre Celebi, Stephen Dusza, David Gutman, Brian Helba, Aadi Kalloo, Konstantinos Liopyris, Michael Marchetti, Harald Kittler, Allan Halpern:,(2018) “Skin Lesion Analysis Toward Melanoma Detection: A Challenge Hosted by the International Skin Imaging Collaboration (ISIC)”, 2018; Online Ref: https://arxiv.org/abs/1902.03368.

  • Tschandl, P., Rosendahl, C. & Kittler, H., (2018) The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions. Sci. Data 5, 180161 DOI:10.1038/sdata.2018.161.

  • Jyothilakshmi K. K and Jeeva J. B, (2014) "Detection of malignant skin diseases based on the lesion segmentation," International Conference on Communication and Signal Processing, DOI: 10.1109/ICCSP.2014.6949867, pp. 382-386.

  • Nitesh Ramakrishnan, Anandhanarayanan Kamalakannan, Balika. J. Chelliah & Govindaraj Rajamanickam, (2019) “Computer Vision Framework for Visual Sharp Object Detection using Deep Learning Model”, International Journal of Engineering and Advanced Technology (IJEAT), ISSN: 2249-8958, Vol-8, No.4, pp. 477-481.


Abstract Views: 337

PDF Views: 166




  • Self-Learning AI Framework for Skin Lesion Image Segmentation and Classification

Abstract Views: 337  |  PDF Views: 166

Authors

Anandhanarayanan Kamalakannan
Department of Cyber Physical Systems, Central Electronics Engineering Research Institute, CSIR Madras Complex, Taramani, Chennai, India
Shiva Shankar Ganesan
Dept. of Computer Science, University of Bridgeport, 126 Park Ave, Bridgeport, United States
Govindaraj Rajamanickam
Department of Cyber Physical Systems, Central Electronics Engineering Research Institute, CSIR Madras Complex, Taramani, Chennai, India

Abstract


Image segmentation and classification are the two main fundamental steps in pattern recognition. To perform medical image segmentation or classification with deep learning models, it requires training on large image dataset with annotation. The dermoscopy images (ISIC archive) considered for this work does not have ground truth information for lesion segmentation. Performing manual labelling on this dataset is time-consuming. To overcome this issue, self-learning annotation scheme was proposed in the two-stage deep learning algorithm. The two-stage deep learning algorithm consists of U-Net segmentation model with the annotation scheme and CNN classifier model. The annotation scheme uses a K-means clustering algorithm along with merging conditions to achieve initial labelling information for training the U-Net model. The classifier models namely ResNet-50 and LeNet-5 were trained and tested on the image dataset without segmentation for comparison and with the U-Net segmentation for implementing the proposed self-learning Artificial Intelligence (AI) framework. The classification results of the proposed AI framework achieved training accuracy of 93.8% and testing accuracy of 82.42% when compared with the two classifier models directly trained on the input images.

Keywords


Self-Learning Annotation Scheme, K-Means Clustering, U-Net, Deep Learning & Skin Lesion Image.

References