Deep Multi-Stage System for Liver and Lesions Segmentation Using CT Images

Authors

Nermeen Elmenabawy, Department of Electronics and Communications Engineering (ECE), Faculty of Engineering, Mansoura University, Mansoura 35516, Egypt
Mervat El-Seddek, Department of Communications and Electronics Engineering, Misr Higher Institute of Engineering and Technology, Mansoura 7651012, Egypt
Hossam El-Din Moustafa, Department of Electronics and Communications Engineering (ECE), Faculty of Engineering, Mansoura University, Mansoura 35516, Egypt
Ahmed Elnakib, Department of Electronics and Communications Engineering (ECE), Faculty of Engineering, Mansoura University, Mansoura 35516, EgyptFollow

Corresponding Author

Ahmed Elnakib

Subject Area

Electronics and Communication Engineering

Article Type

Original Study

Abstract

Accurate tumor segmentation plays a pivotal role in cancer diagnosis and treatment planning. This study introduces an automated system for segmenting liver and tumors in CT images through a multi-stage process. Initial preprocessing improves image quality, followed by feature extraction using pretrained VGG16-Segnet and FCNAlexnet models. Outputs from both models are combined using a parallel fusion operation. Finally, pixel-wise classification designates pixels as liver, tumor, or background. Evaluation on the MICCAI'2017 LiTS database yielded a Dice coefficient of 95.3% for liver and 78.1% for tumors using 5-fold cross-validation. Comparative studies highlight the superior accuracy of our approach in liver and tumor segmentation, promising advancements in clinical diagnosis and treatment strategies.

Keywords

Tumors; Computed Tomography; Deep Learning; Hepatic; Parallel Fusion; Convolution Network

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 License.

Recommended Citation

Elmenabawy, Nermeen; El-Seddek, Mervat; Moustafa, Hossam El-Din; and Elnakib, Ahmed (2023) "Deep Multi-Stage System for Liver and Lesions Segmentation Using CT Images," Mansoura Engineering Journal: Vol. 49 : Iss. 1 , Article 7.
Available at: https://doi.org/10.58491/2735-4202.3128