Melanoma skin cancer identification with amalgamated TSBTC and BTC colour features using ensemble of machine learning algorithms
by Sudeep D. Thepade; Gaurav Ramnani; Shubham Mandhare
International Journal of Computational Vision and Robotics (IJCVR), Vol. 11, No. 6, 2021

Abstract: Manual diagnosis of diseases is time-consuming, subjective and error prone. There is significant scarcity of medical experts in rural areas. Computer assisted diagnosis may help to overcome these challenges. Melanoma skin cancer may become fatal if not detected during its early stages. In the absence of experienced medical professionals, early diagnosis of melanoma may be attempted using machine learning. This paper proposes the melanoma skin cancer identification from dermoscopy skin images by exploring the ensembles of machine learning algorithms using amalgamation of TSBTC and BTC feature extraction methods with various colour spaces. Experimentations conducted with various colour spaces and machine learning algorithms with ensembles resulted in 432 variations of proposed technique. Considering the average of accuracy, sensitivity and specificity; ensemble of AD tree-random forest-SVM in YCbCr colour space with TSBTC features performs best, followed by ensemble of random tree-random forest-AD Tree-SVM in LUV colour space with TSBTC features.

Online publication date: Thu, 28-Oct-2021

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Computational Vision and Robotics (IJCVR):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com