Electrical Engineering and Systems Science > Image and Video Processing
[Submitted on 30 Nov 2019]
Title:Convolutional neural networks model improvements using demographics and image processing filters on chest x-rays
View PDFAbstract:Purpose: The purpose of this study was to observe change in accuracies of convolutional neural networks (CNN) models (ratio of correct classifications to total predictions) on thoracic radiological images by creating different binary classification models based on age, gender, and image pre-processing filters on 14 pathologies.
Methodology: This is a quantitative research exploring variation in CNN model accuracies. Radiological thoracic images were divided by age and gender and pre-processed by various image processing filters.
Findings: We found partial support for enhancement to model accuracies by segregating modeling images by age and gender and applying image processing filters even though image processing filters are sometimes thought of as information filters.
Research limitations: This study may be biased because it is based on radiological images by another research that tagged the images using an automated process that was not checked by a human.
Practical implications: Researchers may want to focus on creating models segregated by demographics and pre-process the modeling images using image processing filters. Practitioners developing assistive technologies for thoracic diagnoses may benefit from incorporating demographics and employing multiple models simultaneously with varying statistical likelihood.
Originality/value: This study uses demographics in model creation and utilizes image processing filters to improve model performance.
Keywords: Convolutional Neural Network (CNN), Chest X-Ray, ChestX-ray14, Lung, Atelectasis, Cardiomegaly, Consolidation, Edema, Effusion, Emphysema, Infiltration, Mass, Nodule, Pleural Thickening, Pneumonia, Pneumathorax
Current browse context:
eess.IV
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.