A lightweight neural network with multiscale feature enhancement for liver CT segmentation

A lightweight neural network with multiscale feature enhancement for liver CT segmentation

<p dir="ltr">Segmentation of abdominal Computed Tomography (CT) scan is essential for analyzing, diagnosing, and treating visceral organ diseases (e.g., hepatocellular carcinoma). This paper proposes a novel neural network (Res-PAC-UNet) that employs a fixed-width residual UNet backb...

Full description

Saved in:
Bibliographic Details
Main Author: Mohammed Yusuf Ansari (16904523) (author)
Other Authors: Yin Yang (35103) (author), Shidin Balakrishnan (14150580) (author), Julien Abinahed (14151792) (author), Abdulla Al-Ansari (14150583) (author), Mohamed Warfa (18282250) (author), Omran Almokdad (18282253) (author), Ali Barah (14777533) (author), Ahmed Omer (18282256) (author), Ajay Vikram Singh (204056) (author), Pramod Kumar Meher (17316988) (author), Jolly Bhadra (14147823) (author), Osama Halabi (14158905) (author), Mohammad Farid Azampour (18282259) (author), Nassir Navab (6254012) (author), Thomas Wendler (13175656) (author), Sarada Prasad Dakua (14151789) (author)
Published: 2022
Subjects:
Biomedical and clinical sciences
Clinical sciences
Engineering
Biomedical engineering
Information and computing sciences
Artificial intelligence
Machine learning
Abdominal Computed Tomography (CT)
Segmentation
Visceral organ diseases
Hepatocellular carcinoma
Neural network
Res-PAC-UNet
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:<p dir="ltr">Segmentation of abdominal Computed Tomography (CT) scan is essential for analyzing, diagnosing, and treating visceral organ diseases (e.g., hepatocellular carcinoma). This paper proposes a novel neural network (Res-PAC-UNet) that employs a fixed-width residual UNet backbone and Pyramid Atrous Convolutions, providing a low disk utilization method for precise liver CT segmentation. The proposed network is trained on medical segmentation decathlon dataset using a modified surface loss function. Additionally, we evaluate its quantitative and qualitative performance; the Res16-PAC-UNet achieves a Dice coefficient of 0.950 ± 0.019 with less than half a million parameters. Alternatively, the Res32-PAC-UNet obtains a Dice coefficient of 0.958 ± 0.015 with an acceptable parameter count of approximately 1.2 million.</p><p dir="ltr">Publisher Correction: A lightweight neural network with multiscale feature enhancement for liver CT segmentation: <a href="https://dx.doi.org/10.1038/s41598-022-20472-5" target="_blank">https://dx.doi.org/10.1038/s41598-022-20472-5</a>, published online 21 September 2022.</p><h2>Other Information</h2><p dir="ltr">Published in: Scientific Reports<br>License: <a href="https://creativecommons.org/licenses/by/4.0" target="_blank">https://creativecommons.org/licenses/by/4.0</a><br>See article on publisher's website: <a href="https://dx.doi.org/10.1038/s41598-022-16828-6" target="_blank">https://dx.doi.org/10.1038/s41598-022-16828-6</a></p>

Similar Items