Volume 8, Issue 1 (6-2021)
jhbmi 2021, 8(1): 67-83 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Nazanin N, Karimi Moridani M, Mahmoudi H. Intelligent Diagnosis of Actinic Keratosis and Squamous Cell Carcinoma of the Skin, Using Linear and Nonlinear Features Based on Image Processing Techniques. jhbmi 2021; 8 (1) :67-83
URL: http://jhbmi.ir/article-1-583-en.html
URL: http://jhbmi.ir/article-1-583-en.html
Ph.D. in Biomedical Engineering, Assistant Professor, Biomedical Engineering Dept., Faculty of Health and Biomedical Engineering, Tehran Medical Sciences, Islamic Azad University, Tehran, Iran
Abstract: (2012 Views)
Introduction: Most skin cancers are treatable in the early stages; thus, an early and rapid diagnosis can be very important to save patients’ lives. Today, with artificial intelligence, early detection of cancer in the initial stages is possible.
Method: In this descriptive-analytical study, a computerized diagnostic system based on image processing techniques was presented, which is much more helpful for the patient. In this method, dermoscopic images of actinic keratosis and squamous cell carcinoma were improved by preprocessing techniques and the potential noises were removed. Then, segmentation was performed using the thresholding method to separate the lesion from the underlying skin. Thereafter, from the segmented area, texture, shape, and color information and features were extracted. Finally, the feature reduction method and support vector machine (SVM) were used to evaluate the proposed method qualitatively and quantitatively.
Results: The data in this study included 100 samples of actinic keratosis images and 100 samples of squamous cell carcinoma. The results of the present study showed that using the genetic algorithm method together with the support vector machine method could help identify the type of skin cancer with 99.7 ± 0.4% accuracy.
Conclusion: The effect of different tissue features in diagnosing the type of lesion showed an increase in the amount and variety of features extracted from the samples would lead to better training and more accurate analysis of the system.
Method: In this descriptive-analytical study, a computerized diagnostic system based on image processing techniques was presented, which is much more helpful for the patient. In this method, dermoscopic images of actinic keratosis and squamous cell carcinoma were improved by preprocessing techniques and the potential noises were removed. Then, segmentation was performed using the thresholding method to separate the lesion from the underlying skin. Thereafter, from the segmented area, texture, shape, and color information and features were extracted. Finally, the feature reduction method and support vector machine (SVM) were used to evaluate the proposed method qualitatively and quantitatively.
Results: The data in this study included 100 samples of actinic keratosis images and 100 samples of squamous cell carcinoma. The results of the present study showed that using the genetic algorithm method together with the support vector machine method could help identify the type of skin cancer with 99.7 ± 0.4% accuracy.
Conclusion: The effect of different tissue features in diagnosing the type of lesion showed an increase in the amount and variety of features extracted from the samples would lead to better training and more accurate analysis of the system.
Keywords: Skin Cancer, Actinic Keratosis, Squamous Cell Carcinoma, Image Processing, Intelligent Diagnosis
Type of Study: Original Article |
Subject:
Artificial Intelligence in Healthcare
Received: 2021/04/17 | Accepted: 2021/04/27
Received: 2021/04/17 | Accepted: 2021/04/27
Audio File [MP3 1477 KB] (42 Download)
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
