An optimized CNN based intelligent prognostics model for disease prediction and classification from Dermoscopy images
- PDF / 9,672,055 Bytes
- 19 Pages / 439.37 x 666.142 pts Page_size
- 91 Downloads / 241 Views
An optimized CNN based intelligent prognostics model for disease prediction and classification from Dermoscopy images Ankita Tyagi 1
& Ritika Mehra
1
Received: 15 June 2019 / Revised: 23 March 2020 / Accepted: 15 May 2020 # Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract
Tiny skin vessels and telangiectasia are most imperative dermoscopy configurations used to differentiate Basal Cell Carcinoma (BCC) from benign skin lesions. This research work builds off of previously developed image analysis techniques to identify vessels automatically to separate benign lesions from BCCs. In this paper, to develop a model for Intelligent Prognostics Model for Disease Prediction and Classification (IPM-DPC) from dermoscopy images is presented using the combination of Convolutional Neural Network (CNN) structure along with the Particle Swarm Optimization (PSO). Here PSO play two different roles in this proposed IPM-DPC, firstly PSO used with K-means segmentation technique to improve the segmentation accuracy then PSO is used as filter for the CNN to train the proposed IPM-DPC. Speed up Robust Features (SURF) algorithm is used as feature descriptor along with PSO as feature selection algorithm which increase the classification accuracy of the system. This study uses a dataset of 1000 dermoscopy skin lesion images of 545 BCCs and 455Non-BCCs or benign images as the input sets. This dataset is taken from ISBI-2016 Dataset and available on:www.isic-archive.com. Experimental results yielded a diagnostic accuracy as high as 99.46% using the IMP-DPC approach, providing a14.94% improvement over a system without using the PSO as filter layer in CNN. When the evaluation parameters of proposed IMP-DPC is compared with a few other state-of-art methods, the proposed method achieves the best performance in terms of accuracy and detection time in differentiating BCC and Non-BCC from dermoscopy skin lesion images.
* Ankita Tyagi [email protected] Ritika Mehra [email protected]
1
School of Computing, DIT University, Dehradun, India
Multimedia Tools and Applications
Keywords Intelligent prognostics model . Skin lesion segmentation . Lesion hair removal . Convolutional neural network (CNN) . Particle swarm optimization (PSO) . Speed up robust features (SURF) descriptor
1 Introduction In Basal Cell Carcinoma (BCC), Telangiectasia is a dilated blood vessels of changeable diameter within the external dermis. These vessels are pragmatic with a number of diseases and are a prominent feature of BCC, which is the most common and dangerous skin cancer [15]. They are generally present as a background skin feature in fair-skinned peoples, particularly in sun-exposed areas in adult peoples. BCC can best be visualized with dermoscopy technique, using either a glass plate with fluid interface which are normally known as the contact non-polarized dermoscopy or cross-polarized lighting which are normally known as the non-contact polarized dermoscopy, together with power exaggeration. In BCC, the cl
Data Loading...
