Edge Enhanced Densely Connected Convolutional Network for Liver Tumor Segmentation

Diagnosing liver sickness is a substantial curative exertion in underdeveloped nations. The previous liver aberration finding approaches have observed a great concept for restraint parameters but low in accuracy. The abrasion on the liver hasn't been notorious obviously with previous representa...

Full description

Saved in:
Bibliographic Details
Main Authors: Kumari, L. Krishna, Anusuya, V., Ramalakshmi, K., Rajalakshmi, R, Theivanathan, G., Lakshmi, A.
Format: Conference Proceeding
Language:English
Subjects:
Accuracy
Communication systems
Convolutional neural networks
Densely Connected Convolutional Networks
Filtering
Image edge detection
Liver
Liver abnormality
Liver abnormality detection
Measurement
Median filtering
Neural networks
Training
Tumors
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Diagnosing liver sickness is a substantial curative exertion in underdeveloped nations. The previous liver aberration finding approaches have observed a great concept for restraint parameters but low in accuracy. The abrasion on the liver hasn't been notorious obviously with previous representations, so progressive, effectual, and genuine liver sickness discovery is vital. To overwhelm the boundaries of traditional approaches, this method suggests deep liver irregularity discovery with neural networks having a thick connection called as Densely Connected Convolutional Networks (DCCN). The pre-processing has been accomplished through median filtering, and training is executed over DCCN conventional methods with the performance metrics of 92.34 % accuracy, 91.54 %, Precision and 93.65 % of recall.
ISSN:2469-5556
DOI:10.1109/ICACCS60874.2024.10717318