A Novel Nodule Tracking System and Convolutional Neural Networks for Medical Internet of Things

Medical and health care is a significant application field for Internet of things technologies. The medical Internet of Things (IoT) is employed in this research to collect data from clinical trials of cancerous and non-cancerous lung nodules, and intelligent diagnostic strategies are studied as a r...

Full description

Saved in:
Bibliographic Details
Published in2022 4th IEEE Middle East and North Africa COMMunications Conference (MENACOMM) pp. 197 - 202
Main Authors Bejaoui, Dhia Eddine, Mahersia, Hela, Bejaoui, Tarek
Format Conference Proceeding
LanguageEnglish
Published IEEE 06.12.2022
Subjects
CNN
Computed tomography
CT Scan
Image databases
Image edge detection
Image segmentation
Internet of Things
Lung
Lung cancer
Medical services
Nodules
Segmentation
Tracking
Online AccessGet full text

Cover

More Information
Summary:Medical and health care is a significant application field for Internet of things technologies. The medical Internet of Things (IoT) is employed in this research to collect data from clinical trials of cancerous and non-cancerous lung nodules, and intelligent diagnostic strategies are studied as a result. Detecting cancer in its early stages offers the patient a highly improved survival rate and various treatment options. Oncologists use CT scans as one of the primary screening techniques to aid in cancer cell identification. This research aims to establish a new methodology to extract and detect nodules residing in lung regions using image processing and tracking techniques. Our proposed system starts with segmenting lung regions by detecting and processing each side individually for efficient preservation of suspicious nodules located on the edge of the lung parenchyma. This is followed by tracking and listing every candidate that occurs in the screening image of the patient to extract highly suspicious nodules that pass finally through two CNN models to distinguish True Nodules. The proposed method gives a high accuracy rate compared to other methods (around 95%). That is why it is one of the recommendations for efficient outcomes. Our database in this approach is the Lung Image Database Consortium image collection (LIDC-IDRI).
DOI:10.1109/MENACOMM57252.2022.9998280