Downloads: 2 | Views: 157 | Weekly Hits: ⮙2 | Monthly Hits: ⮙2

Research Paper | Computer Science and Information Technology | India | Volume 12 Issue 10, October 2023 | Popularity: 5.2 / 10

Rate This Article!      

Modeling of Dragonfly Algorithm with Deep Learning for Skin Cancer Diagnosis on Dermoscopic Images

Vijay Arumugam R., Saravanan S.

Abstract: Skin cancer, a widespread and possibly life-threatening condition, requires an earlier and accurate diagnosis for efficient involvement. Dermoscopic images, providing a window into skin lesions, gives a useful resource for medical specialists in this context. This article goals to develop a complex architecture to detect skin cancer employing dermatoscopy images, integrating cutting-edge techniques in image analysis, machine learning, and artificial intelligence. This manuscript introduces the Dragonfly Algorithm with Deep Learning for Skin Cancer Diagnoses on Dermoscopy Images (DFADL-SCDDI) method. Our technique integrates advanced methods namely feature extraction, preprocessing, classification, and parameter optimization to increase the reliability and of accuracy identification. For image preprocessing, we exploit the Gabor filter (GF), a robust tool for enriching texture and structure data in images. Feature extraction has been executed employing a Capsule Network (CapsNet). CapsNet is a deep learning (DL) model that exceeds in capturing hierarchical and in-depth features in images. Classification is conducted by a Gated Recurrent Unit (GRU), a kind of recurrent neural network (RNN) ability to model sequential patterns and dependencies within the feature representations. To additional improve the model's effectiveness; we implement the Dragonfly Algorithm (DA) for parameter tuning. The DA has a powerful optimization system stimulated by nature, developed for enhancing hyperparameters efficiently, consequently higher the model's diagnostic accuracy. The proposed architecture is assessed on a large database of dermatoscopy images, signifying its effectiveness in skin cancer detection. The outcomes exhibit substantial enhancement in reliability and accuracy compared to traditional systems.

Keywords: Dragonfly Algorithm; Dermoscopy; Gated Recurrent Unit; Deep Learning; Parameter Tuning

Edition: Volume 12 Issue 10, October 2023

Pages: 1916 - 1922

DOI: https://www.doi.org/10.21275/SR231025120749

Make Sure to Disable the Pop-Up Blocker of Web Browser

Click to Cite this Article

Similar Articles