Home / Current Issue / Paper 1716185
Liver Tumor Detection Using Deep Learning
Subject area: Science,Engineering and Technology · Area of research: Medical Imaging, and Healthcare Technology
DOI: https://doi.org/10.64388/IREV9I10-1716185
Abstract
Liver cancer remains one of the most serious health challenges across the world, with a high mortality rate largely due to late diagnosis. Detecting tumors at an early stage is critical for improving survival chances, yet traditional methods depend heavily on manual analysis of medical images. This process is not only time-consuming but also subject to human error and variation between specialists. With the rise of artificial intelligence, especially deep learning, there has been a significant shift toward automating the solutions in medical imaging. These technologies offer faster and more consistent analysis, making them highly valuable in clinical settings. Among various imaging techniques, computed tomography (CT) scans are widely used because they provide detailed views of liver structures and abnormalities. In this research, a deep learning-based approach is proposed for detecting liver tumors from CT images. The model is built using a 3D U-Net architecture, which is well-suited for capturing spatial relationships in volumetric data. Unlike traditional 2D models, the 3D approach analyzes multiple image slices together, leading to more accurate segmentation of tumor regions. One of the key challenges in deep learning models is selecting the right hyperparameters, such as learning rate and batch size. Instead of relying on manual tuning, this study incorporates the Bat Algorithm, a nature-inspired optimization technique. This method helps automatically find the best parameter settings, improving both the efficiency and performance of the model. The dataset used in this work includes annotated CT images, which are preprocessed through normalization, resizing, and augmentation. These steps ensure better data quality and help the model generalize well to new cases. Experimental results show that the proposed method achieves high accuracy and reliable performance in tumor detection. It also maintains a good balance between identifying tumors and avoiding false detections. Overall, this study demonstrates how combining deep learning with optimization techniques can significantly enhance liver cancer detection. The approach has strong potential to support doctors in making faster and more accurate diagnoses, ultimately improving patient care.
References
[1] J. Wang, Y. Peng, S. Jing, L. Han, T. Li, and J. Luo, ―A deep-learning approach for segmentation of liver tumors in magnetic resonance imaging using UNet++ (Periodical style),‖ BMC Cancer, vol. 23, no. 1, pp. 1060, Nov. 2023.
[2] I. Manghi, A. Pecchi, G. Esposito, M. Macheda, E. Zanni, F. Di Benedetto, and F. Ferraguti, ―Deep learning for automatic segmentation of hepatocellular carcinoma in contrast enhanced CT scans (Periodical style),‖ Scientific Reports, vol. 15, no. 1, pp. 42000, Nov. 2025.
[3] Z. Song, H. Wu, W. Chen, and A. Slowik, ―Improving automatic segmentation of liver tumor images using a deep learning model (Periodical style),‖ Heliyon, vol. 10, no. 7, e28538, Apr. 2024.
[4] C. Wu, Q. Chen, H. Wang, Y. Guan, Z. Mian, C. Ruan, Q. Song, H. Jiang, J. Pan, and X. Li, ―A review of deep learning approaches for multimodal image segmentation of liver cancer (Periodical style),‖ J. Appl. Clin. Med. Phys., vol. 25, no. 12, e14540, Oct. 2024.
[5] I.-C. Lee, Y.-P. Tsai, Y.-C. Lin, T.-C. Chen, C.-H. Yen, N.-C. Chiu, H.-E. Hwang, C.-A. Liu, J.-G. Huang, R.-C. Lee, and Y.-H. Huang, ―A hierarchical fusion strategy of deep learning networks for detection and segmentation of hepatocellular carcinoma from CT images (Periodical style),‖ Cancer Imaging, vol. 24, article 43, Mar. 2024.
[6] D. Wei, Y. Jiang, X. Zhou, D. Wu, and X. Feng, ―A review of advancements and challenges in liver segmentation (Periodical style),‖ J. Imaging, vol. 10, no. 8, pp. 202, Aug. 2024.
How to cite this paper
@article{1716185,
author = {Soniya Komal V, Bajarang S, Darshan N, Afjal K, Faizan},
title = {Liver Tumor Detection Using Deep Learning},
journal = {Iconic Research And Engineering Journals},
year = {2026},
volume = {9},
number = {10},
pages = {1025-1030},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1716185.pdf},
abstract = {Liver cancer remains one of the most serious health challenges across the world, with a high mortality rate largely due to late diagnosis. Detecting tumors at an early stage is critical for improving survival chances, yet traditional methods depend heavily on manual analysis of medical images. This process is not only time-consuming but also subject to human error and variation between specialists. With the rise of artificial intelligence, especially deep learning, there has been a significant shift toward automating the solutions in medical imaging. These technologies offer faster and more consistent analysis, making them highly valuable in clinical settings. Among various imaging techniques, computed tomography (CT) scans are widely used because they provide detailed views of liver structures and abnormalities. In this research, a deep learning-based approach is proposed for detecting liver tumors from CT images. The model is built using a 3D U-Net architecture, which is well-suited for capturing spatial relationships in volumetric data. Unlike traditional 2D models, the 3D approach analyzes multiple image slices together, leading to more accurate segmentation of tumor regions. One of the key challenges in deep learning models is selecting the right hyperparameters, such as learning rate and batch size. Instead of relying on manual tuning, this study incorporates the Bat Algorithm, a nature-inspired optimization technique. This method helps automatically find the best parameter settings, improving both the efficiency and performance of the model. The dataset used in this work includes annotated CT images, which are preprocessed through normalization, resizing, and augmentation. These steps ensure better data quality and help the model generalize well to new cases. Experimental results show that the proposed method achieves high accuracy and reliable performance in tumor detection. It also maintains a good balance between identifying tumors and avoiding false detections. Overall, this study demonstrates how combining deep learning with optimization techniques can significantly enhance liver cancer detection. The approach has strong potential to support doctors in making faster and more accurate diagnoses, ultimately improving patient care.},
month = {April},
doi = {https://doi.org/10.64388/IREV9I10-1716185}
}