Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{187852,
        author = {Shaik Roshna and Chiranjeevi Althi and Dr. T. Ramashri},
        title = {An Enhanced Attention U-Net++ Architecture for Accurate Liver and Tumor Segmentation in MRI Images},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {12},
        number = {6},
        pages = {7400-7405},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=187852},
        abstract = {Computer-aided diagnosis and treatment planning require the proper liver and liver tumor segmentation in the magnetic resonance image (MRI) scans. However, the task is difficult due to irregular tumor boundaries, insufficient contrast between tumor and healthy tissues, and inconsistency in the quality of images. This study proposes a deep learning-based multi-label Attention U-Net++ architecture for liver and tumor segmentation. The first part of the workflow is a preprocessing pipeline which involves resizing, normalization, denoising and contrast enhancement to improve the image quality. Attention U-Net++ model incorporates skip connections and attention processes within a model to focus on the areas of tumor and maintain the liver boundaries. The presented framework was trained and tested on ATLAS MRI data with a Dice score of 0.93 on liver and 0.85 on tumor with an average accuracy of 99%. These findings prove that preprocessing and Attention and U-Net++ work together to greatly enhance segmentation performance and has good prospects in the computer-aided diagnosis and treatment planning.},
        keywords = {Attention U-Net++, Deep Learning, Liver Tumor Segmentation, Medical Image Analysis, MRI.},
        month = {December},
        }