Estimation of level of liver damage due to cancer using deep convolutional neural network in CT images

1Citations
Citations of this article
10Readers
Mendeley users who have this article in their library.
Get full text

Abstract

The lesion size estimation is essential need while diagnosing the liver cancer and treatment scenario. The lesion segmentation suing conventional methods such as region growing, threshold based segmentation provide limited performance due to variations in light intensity distribution throughout the image. The deep learning approach used in this paper consist of input dataset of liver abdominal images along with labelled set combination of variety of liver regions and lesion structures. The care has been taken while constructing the dataset such that, the lesion due to cancer in liver of particular image should have at least one matching structure should be present in one of the labelled images. The 3 fold validation is done to evaluate the performance in which total 140 images of liver cancer are used for training, 30 images for validation and 30 images for testing. The result shows 98.5% accuracy for lesion classification. The area of lesion is compared to total area of liver in terms of pixels to estimate the total area occupied by the lesion and amount of liver damage.

Cite

CITATION STYLE

APA

Vanmore, S. V., & Chougule, S. R. (2019). Estimation of level of liver damage due to cancer using deep convolutional neural network in CT images. International Journal of Innovative Technology and Exploring Engineering, 9(1), 3761–3764. https://doi.org/10.35940/ijitee.A4818.119119

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free