Severity Gradation of Psoriatic Plaques using Ensemble of Deep Convolutional Neural Networks.

Date of Submission

December 2018

Date of Award

Winter 12-12-2019

Institute Name (Publisher)

Indian Statistical Institute

Document Type

Master's Dissertation

Degree Name

Master of Technology

Subject Name

Computer Science


Computer Vision and Pattern Recognition Unit (CVPR-Kolkata)


Garain, Utpal (CVPR-Kolkata; ISI)

Abstract (Summary of the Work)

Severity gradation of psoriatic plaque is important for the estimation of Psoriasis Area Severity Index (abbreviated as PASI) that facilitates the diagnosis as well as the treatment of the disease. Severity assessment by manual examination of the plaques of the diseased person or by observation of images of the affected skin area suffers from inter and intra-observer variability.Therefore, automated techniques can not only lead to reduced effort but also can reduce inaccuracy, provided a sufficient amount of correctly annotated data is available. Recent advancements of deep learning in computer vision and medical imaging domain has led to a substantial improvement of performance over traditional image processing techniques.This work has proposed five new methods for severity scoring in order to bring about improvement in accuracy from the baseline. The first method uses a small-sized CNN with very less number of parameters for classification; the second method uses a two-stage classification approach based on CNN; the third method is a pair wise CNN based classification method; the fourth one is a majority voting based CNN ensemble and fifth one is a stacking or super-learning based CNN ensemble. Other approaches like Texture CNN based classification have also been tried and depicted in the thesis as well.


ProQuest Collection ID:

Control Number


Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.


This document is currently not available here.