Automated approach for skin lesion segmentation utilizing a hybrid deep learning algorithm

In computer vision segmenting a digital image into multiple segments is a common objective for which convolutional neural networks have been proven to be consistent. Skin lesion segmentation is an important process as it focuses on the specific parts of the skin and it improves manual diagnostics. S...

Full description

Saved in:
Bibliographic Details
Published in:Multimedia tools and applications 2024-05, Vol.83 (15), p.46017-46035
Main Authors: Manjunath, R V, N, Yashaswini Gowda
Format: Article
Language:English
Subjects:
Algorithms
Artificial neural networks
Automation
Computer Communication Networks
Computer Science
Computer vision
Data Structures and Information Theory
Datasets
Deep learning
Digital computers
Digital imaging
Lesions
Machine learning
Multimedia Information Systems
Special Purpose and Application-Based Systems
Track 2: Medical Applications of Multimedia
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In computer vision segmenting a digital image into multiple segments is a common objective for which convolutional neural networks have been proven to be consistent. Skin lesion segmentation is an important process as it focuses on the specific parts of the skin and it improves manual diagnostics. Skin lesion segmentation is always a challenging task due to lesion variation in size, color, boundary, the low contrast between lesion and normal skin and even sometimes inaccurate lesion location detection. To overcome many problems, it is very much essential to develop an automatic skin lesion segmentation algorithm. Here we are investigating using hybrid ResUNet architecture which is a deep learning algorithm. The ISIC 2017 and ISIC 2018 dataset consists of RGB skin lesion images and their binary ground truths of different size. We evaluated our hybrid model for size 128x128 with DSC, Jaccard Index and Accuracy parameters. The results show that hybrid architecture achieves a DSC of 92.61%, a Jaccard Index of 89.93% and it outperforms the other approaches in accuracy with a score of 98.86% for ISIC 2018 dataset. The results are encouraging and can lead to fully-fledged automated approaches for skin lesion segmentation. The comprehensive experimental results demonstrate the efficiency of the proposed approach in the task of skin lesion segmentation.
ISSN:1573-7721
1380-7501
1573-7721
DOI:10.1007/s11042-023-16934-1