Lung Nodule Segmentation Accuracy in CT Images Using YOLO, 3D-CNN, and Ensemble ViT-UNETR U-Net

Reyga Ferdiansyah Putra; Antoni Wibowo; Dewi Retno Sari Saputro

doi:10.59261/jequi.v8i2.308

Authors

Reyga Ferdiansyah Putra Universitas Bina Nusantara
Antoni Wibowo Universitas Bina Nusantara
Dewi Retno Sari Saputro Universitas Sebelas Maret

DOI:

https://doi.org/10.59261/jequi.v8i2.308

Keywords:

Lung Nodules, Hybrid Segmentation, YOLOv12, 3D-CNN, ViT-UNETR, U-Net, Ensemble Learning

Abstract

Background: Lung cancer is the leading cause of cancer-related mortality globally, with over 2.2 million new cases and 1.8 million deaths reported annually (WHO, 2022). Pulmonary nodule detection through low-dose computed tomography (LDCT) screening is the most effective method for early lung cancer identification. However, automated systems still face significant challenges: high false positive rates, limited sensitivity for micronodules (<5 mm), and poor segmentation accuracy for nodules with irregular morphology or juxtapleural attachment.

Objective: Lung nodules early discovery is key to treating lung carcinoma, but even conventional systems' micronodules still have high false positives and low accuracy.

Method: This study presents an end-to-end hybrid pipeline that uses the LUNA16 database to tackle this issue. The initial stage is to make use of YOLOv12 for Region of Interest (ROI) extraction, with 3D-CNN carrying out false positive filtering through volumetric verification as a gate. The final phase conducts pixel-level precision segmentation using Adaptive Bayesian Fusion on U-Net Residual 3D ensemble (local texture features) and ViT-UNETR (global anatomical context).

Results: Experiments showed superior performance level 99.99% Accuracy, Mean Dice Similarity Coefficient (DSC) at 93.88% and IoU is 90.45%. The system was very robust, reaching 97.33% DSC in the micro nodule category (<5 mm).

Conclusion: In summary, this integrated architecture delivers an objective, efficient and high-quality solution for automated Diagnosis.

Downloads

Download data is not yet available.

References

Annavarapu, C. S. R., Parisapogu, S. A. B., Keetha, N. V., Donta, P. K., & Rajita, G. (2023). A Bi-FPN-Based Encoder–Decoder Model for Lung Nodule Image Segmentation. Diagnostics, 13(8), 1406. https://doi.org/10.3390/diagnostics13081406

Asha, V., & Bhavanishankar, K. (2024). Advanced Lung Nodule Segmentation and Classification for Early Detection of Lung Cancer using SAM and Transfer Learning. Preprint, 1–25.

Banu, S. F., Sarker, M. M. K., Abdel-Nasser, M., Puig, D., & Raswan, H. A. (2021). AWEU-Net: an attention-aware weight excitation U-Net for lung nodule segmentation. Applied Sciences, 11(21), 10132.

Bhattacharyya, D., Thirupathi Rao, N., Joshua, E. S. N., & Hu, Y.-C. (2023). A bi-directional deep learning architecture for lung nodule semantic segmentation. The Visual Computer, 39(11), 5245–5261.

Delfan, N., Moghaddam, H. A., Modaresi, M., Afshari, K., Nezamabadi, K., Pak, N., Ghaemi, O., & Forouzanfar, M. (2022). CT-LungNet: A deep learning framework for precise lung tissue segmentation in 3D thoracic CT scans. ArXiv Preprint ArXiv:2212.13971.

Dutande, P., Baid, U., & Talbar, S. (2021). LNCDS: A 2D-3D cascaded CNN approach for lung nodule classification, detection and segmentation. Biomedical Signal Processing and Control, 67, 102527. https://doi.org/10.1016/j.bspc.2021.102527

Gu, X., Zhu, Y., Li, C., Xu, X., Jin, K., & Xu, L. (2025). ShapeField-lung: continuous shape embedding for early lung cancer detection via pulmonary nodule segmentation. Npj Digital Medicine, 8(1), 736. https://doi.org/10.1038/s41746-025-02041-y

Kadia, D. D., Alom, M. Z., Burada, R., Nguyen, T. V., & Asari, V. K. (2021). R 2 U3D: Recurrent Residual 3D U-Net for Lung Segmentation. IEEE Access, 9, 88835–88843. https://doi.org/10.1109/ACCESS.2021.3089704

Li, R., & Honarvar Shakibaei Asli, B. (2026). Multi-Task Deep Learning for Lung Nodule Detection and Segmentation in CT Scans. Electronics, 15(4), 736. https://doi.org/10.3390/electronics15040736

Liu, W., Zhang, L., Li, X., Liu, H., Feng, M., & Li, Y. (2025). A semisupervised knowledge distillation model for lung nodule segmentation. Scientific Reports, 15(1), 10562. https://doi.org/10.1038/s41598-025-94132-9

Mamun, T. B., Madhuri, A., Sobir, N., & Hasan, T. (2025). LMLCC-Net: A Semi-Supervised Deep Learning Model for Lung Nodule Malignancy Prediction from CT Scans using a Novel Hounsfield Unit-Based Intensity Filtering. ArXiv Preprint ArXiv:2505.06370.

Ma, X., Song, H., Jia, X., & Wang, Z. (2024). An improved V-Net lung nodule segmentation model based on pixel threshold separation and attention mechanism. Scientific Reports, 14(1), 4743.

Naseer, I., Masood, T., Akram, S., Ali, Z., Ahmad, A., Rehman, S. U., & Jaffar, A. (2024). Empowering Diagnosis: Cutting-Edge Segmentation and Classification in Lung Cancer Analysis. Computers, Materials & Continua, 79(3), 4963–4977. https://doi.org/10.32604/cmc.2024.050204

Ramezani, H., Vedrines, C., Aleman, D., & Létourneau, D. (2025). LNTransformer: Lung Nodule Transformer for Sparse CT Segmentation. 2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 4238–4245. https://doi.org/10.1109/CVPRW67362.2025.00407

Razlighi, Y. A., Kamali-Asl, A., & Arabi, H. (2022). A hierarchical approach for pulmonary nodules identification from ct images using yolo v5s nodule detection and 3d neural network classifier. ArXiv Preprint ArXiv:2212.09366.

Shuvo, S. B., & Mamun, T. B. (2026). AutoLungDx: A hybrid deep learning approach for early lung cancer diagnosis using 3D Res-U-Net, YOLOv5, and vision transformers. Informatics in Medicine Unlocked, 101739.

Siegel, R. L., Miller, K. D., Wagle, N. S., & Jemal, A. (2023). Cancer statistics, 2023. CA: A Cancer Journal for Clinicians, 73(1), 17–48. https://doi.org/10.3322/caac.21763

Sung, H., Ferlay, J., Siegel, R. L., Laversanne, M., Soerjomataram, I., Jemal, A., & Bray, F. (2021). Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA: A Cancer Journal for Clinicians, 71(3), 209–249. https://doi.org/10.3322/caac.21660

Team, N. L. S. T. R. (2019). Lung cancer incidence and mortality with extended follow-up in the National Lung Screening Trial. Journal of Thoracic Oncology, 14(10), 1732–1742.

Turjya, S. M., & Fawakherji, M. (2026). Federated lung nodule segmentation using a hybrid transformer–U-Net architecture. Scientific Reports, 16(1), 5228. https://doi.org/10.1038/s41598-026-35243-9

Wu, Y., Liu, X., Shi, Y., Chen, X., Wang, Z., Xu, Y., & Wang, S. (2025). S 3 TU-Net: Structured convolution and superpixel transformer for lung nodule segmentation. Medical & Biological Engineering & Computing, 63(12), 3777–3791.

Xiao, Z., Liu, B., Geng, L., Zhang, F., & Liu, Y. (2020). Segmentation of Lung Nodules Using Improved 3D-UNet Neural Network. Symmetry, 12(11), 1787. https://doi.org/10.3390/sym12111787

Zhang, J., Yang, M., Guo, W., Xavier, B. A., Bolen, M., & Li, X. (2025). Detection-guided deep learning-based model with spatial regularization for lung nodule segmentation. Quantitative Imaging in Medicine and Surgery, 15(5), 4204–4216. https://doi.org/10.21037/qims-2024-2511

Zhang, X., Fei, L., & Gong, Q. (2023). A semantic segmentation of the lung nodules using a shape attention-guided contextual residual network. Physics in Medicine & Biology, 68(16), 165017. https://doi.org/10.1088/1361-6560/ace09d

Zhou, Z., Gou, F., Tan, Y., & Wu, J. (2022). A Cascaded Multi-Stage Framework for Automatic Detection and Segmentation of Pulmonary Nodules in Developing Countries. IEEE Journal of Biomedical and Health Informatics, 26(11), 5619–5630. https://doi.org/10.1109/JBHI.2022.3198509