G-UNETR++: A Gradient-Enhanced Network for Accurate and Robust Liver Segmentation from Computed Tomography Images
Accurate liver segmentation from computed tomography (CT) scans is essential for liver cancer diagnosis and liver surgery planning. Convolutional neural network (CNN)-based models have limited segmentation performance due to their localized receptive fields. Hybrid models incorporating CNNs and tran...
Saved in:
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-01-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/15/2/837 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832589185938096128 |
|---|---|
| author | Seungyoo Lee Kyujin Han Hangyeul Shin Harin Park Seunghyon Kim Jeonghun Kim Xiaopeng Yang Jae Do Yang Hee Chul Yu Heecheon You |
| author_facet | Seungyoo Lee Kyujin Han Hangyeul Shin Harin Park Seunghyon Kim Jeonghun Kim Xiaopeng Yang Jae Do Yang Hee Chul Yu Heecheon You |
| author_sort | Seungyoo Lee |
| collection | DOAJ |
| description | Accurate liver segmentation from computed tomography (CT) scans is essential for liver cancer diagnosis and liver surgery planning. Convolutional neural network (CNN)-based models have limited segmentation performance due to their localized receptive fields. Hybrid models incorporating CNNs and transformers that can capture long-range dependencies have shown promising performance in liver segmentation with the cost of high model complexity. Therefore, a new network architecture named G-UNETR++ is proposed to improve accuracy in liver segmentation with moderate model complexity. Two gradient-based encoders that take the second-order partial derivatives (the first two elements from the last column of the Hessian matrix of a CT scan) as inputs are proposed to learn the 3D geometric features such as the boundaries between different organs and tissues. In addition, a hybrid loss function that combines dice loss, cross-entropy loss, and Hausdorff distance loss is designed to address class imbalance and improve segmentation performance in challenging cases. The proposed method was evaluated on three public datasets, the Liver Tumor Segmentation (LiTS) dataset, the 3D Image Reconstruction for Comparison of Algorithms Database (3D-IRCADb), and the Segmentation of the Liver Competition 2007 (Sliver07) dataset, and achieved 97.38%, 97.50%, and 97.32% in terms of the dice similarity coefficient for liver segmentation on the three datasets, respectively. The proposed method outperformed the other state-of-the-art models on the three datasets, which demonstrated the strong effectiveness, robustness, and generalizability of the proposed method in liver segmentation. |
| format | Article |
| id | doaj-art-60b86e2ba1c644758cd7acb96660729c |
| institution | Kabale University |
| issn | 2076-3417 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj-art-60b86e2ba1c644758cd7acb96660729c2025-01-24T13:21:00ZengMDPI AGApplied Sciences2076-34172025-01-0115283710.3390/app15020837G-UNETR++: A Gradient-Enhanced Network for Accurate and Robust Liver Segmentation from Computed Tomography ImagesSeungyoo Lee0Kyujin Han1Hangyeul Shin2Harin Park3Seunghyon Kim4Jeonghun Kim5Xiaopeng Yang6Jae Do Yang7Hee Chul Yu8Heecheon You9School of Global Entrepreneurship and Information Communication Technology, Handong Global University, Pohang 37554, Republic of KoreaSchool of Global Entrepreneurship and Information Communication Technology, Handong Global University, Pohang 37554, Republic of KoreaSchool of Global Entrepreneurship and Information Communication Technology, Handong Global University, Pohang 37554, Republic of KoreaSchool of Global Entrepreneurship and Information Communication Technology, Handong Global University, Pohang 37554, Republic of KoreaSchool of Global Entrepreneurship and Information Communication Technology, Handong Global University, Pohang 37554, Republic of KoreaSchool of Global Entrepreneurship and Information Communication Technology, Handong Global University, Pohang 37554, Republic of KoreaSchool of Global Entrepreneurship and Information Communication Technology, Handong Global University, Pohang 37554, Republic of KoreaDepartment of Surgery, Jeonbuk National University Medical School and Hospital, Jeonju 54907, Republic of KoreaDepartment of Surgery, Jeonbuk National University Medical School and Hospital, Jeonju 54907, Republic of KoreaDepartment of Industrial Management and Engineering, Pohang University of Science and Technology, Pohang 37673, Republic of KoreaAccurate liver segmentation from computed tomography (CT) scans is essential for liver cancer diagnosis and liver surgery planning. Convolutional neural network (CNN)-based models have limited segmentation performance due to their localized receptive fields. Hybrid models incorporating CNNs and transformers that can capture long-range dependencies have shown promising performance in liver segmentation with the cost of high model complexity. Therefore, a new network architecture named G-UNETR++ is proposed to improve accuracy in liver segmentation with moderate model complexity. Two gradient-based encoders that take the second-order partial derivatives (the first two elements from the last column of the Hessian matrix of a CT scan) as inputs are proposed to learn the 3D geometric features such as the boundaries between different organs and tissues. In addition, a hybrid loss function that combines dice loss, cross-entropy loss, and Hausdorff distance loss is designed to address class imbalance and improve segmentation performance in challenging cases. The proposed method was evaluated on three public datasets, the Liver Tumor Segmentation (LiTS) dataset, the 3D Image Reconstruction for Comparison of Algorithms Database (3D-IRCADb), and the Segmentation of the Liver Competition 2007 (Sliver07) dataset, and achieved 97.38%, 97.50%, and 97.32% in terms of the dice similarity coefficient for liver segmentation on the three datasets, respectively. The proposed method outperformed the other state-of-the-art models on the three datasets, which demonstrated the strong effectiveness, robustness, and generalizability of the proposed method in liver segmentation.https://www.mdpi.com/2076-3417/15/2/837liver segmentationgradient-enhanced networkdeep learningtransformer |
| spellingShingle | Seungyoo Lee Kyujin Han Hangyeul Shin Harin Park Seunghyon Kim Jeonghun Kim Xiaopeng Yang Jae Do Yang Hee Chul Yu Heecheon You G-UNETR++: A Gradient-Enhanced Network for Accurate and Robust Liver Segmentation from Computed Tomography Images Applied Sciences liver segmentation gradient-enhanced network deep learning transformer |
| title | G-UNETR++: A Gradient-Enhanced Network for Accurate and Robust Liver Segmentation from Computed Tomography Images |
| title_full | G-UNETR++: A Gradient-Enhanced Network for Accurate and Robust Liver Segmentation from Computed Tomography Images |
| title_fullStr | G-UNETR++: A Gradient-Enhanced Network for Accurate and Robust Liver Segmentation from Computed Tomography Images |
| title_full_unstemmed | G-UNETR++: A Gradient-Enhanced Network for Accurate and Robust Liver Segmentation from Computed Tomography Images |
| title_short | G-UNETR++: A Gradient-Enhanced Network for Accurate and Robust Liver Segmentation from Computed Tomography Images |
| title_sort | g unetr a gradient enhanced network for accurate and robust liver segmentation from computed tomography images |
| topic | liver segmentation gradient-enhanced network deep learning transformer |
| url | https://www.mdpi.com/2076-3417/15/2/837 |
| work_keys_str_mv | AT seungyoolee gunetragradientenhancednetworkforaccurateandrobustliversegmentationfromcomputedtomographyimages AT kyujinhan gunetragradientenhancednetworkforaccurateandrobustliversegmentationfromcomputedtomographyimages AT hangyeulshin gunetragradientenhancednetworkforaccurateandrobustliversegmentationfromcomputedtomographyimages AT harinpark gunetragradientenhancednetworkforaccurateandrobustliversegmentationfromcomputedtomographyimages AT seunghyonkim gunetragradientenhancednetworkforaccurateandrobustliversegmentationfromcomputedtomographyimages AT jeonghunkim gunetragradientenhancednetworkforaccurateandrobustliversegmentationfromcomputedtomographyimages AT xiaopengyang gunetragradientenhancednetworkforaccurateandrobustliversegmentationfromcomputedtomographyimages AT jaedoyang gunetragradientenhancednetworkforaccurateandrobustliversegmentationfromcomputedtomographyimages AT heechulyu gunetragradientenhancednetworkforaccurateandrobustliversegmentationfromcomputedtomographyimages AT heecheonyou gunetragradientenhancednetworkforaccurateandrobustliversegmentationfromcomputedtomographyimages |