Integrating Metaheuristics in Computer Vision for Real-World Optimization Problems

19A Thyroid Nodule Detection Using L1-Norm Inception Deep Neural Network

Saranya G.

Department of Biomedical Engineering, Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology, Avadi, Chennai, India

Abstract

In the current scenario, various computer-assist diagnosis models are available to detect the thyroid nodule. The complication with the model is to correctly identify whether the affected nodule is normal or abnormal. At present, echo (sound wave) is advised to do the initial and timely process to diagnosis the thyroid nodules. But, attaining an acceptable prognosis from ultrasound image hangs on the radiologist skill and further circumstance. Prodigious pressure is put on the automated version and for the safe mechanism to monitor the ultrasound images effectual. The Proposed work gives a stacking methodology for automatic thyroid nodule classification using L1-norm inception deep neural network (L1-IDNN). The medical ultrasound image is used to perform the fusion process, and it is carried out to strengthen the image by applying adaptive histogram equalization technique and variance stretching. The thyroid nodule images are fused with the help of hybrid pyramid fusion algorithm. The result claims that the initiated model provides preferable values in respect of accuracy compared to the normal machine learning technique.

Keywords: Thyroid nodules, ultrasonography, L1-IDNN, hybrid pyramid fusion, stacking method

19.1 Introduction

The serious cancer ...