SavedAdd your journal

Journals / Erciyes Üniversitesi Fen Bilimleri Enstitüsü Fen Bilimleri Dergisi / 2026 / Cilt: 42 Sayı: 2

A Multi-Stage Pipeline for 3D Reconstruction of Vehicles from Heterogeneous Multi-View Images

Heterojen Çoklu Görüntülerden Araçların 3B Yeniden Oluşturulması için Çok Aşamalı Bir Yaklaşım

Muzaffer Arda Uslu(ESKİŞEHİR TEKNİK ÜNİVERSİTESİ)0009-0009-0537-7003
Selcan Kaplan Berkaya(ESKİŞEHİR TEKNİK ÜNİVERSİTESİ)0000-0001-6728-4050
Pages
—
Publication Date
2026-06-26
🏛️Bibliographic metadata and full text access for this article are provided via DergiPark.
DergiPark Record ↗
✨

AI Summary & Key Takeaways

AI Insight
⚡ TL;DR(Too Long; Didn't Read)

This contribution establishes a standardized benchmark resource for tasks such as novel view synthesis, damage assessment, and vehicle appearance modeling, representing one of the first publicly available datasets of vehicle-class-specific 3D Gaussian Splatting reconstructions derived from heterogeneous multi-source collections.

📌 3 Key Takeaways

🎯
Objective & Hypothesis:Background/Aim: Three-dimensional (3D) reconstruction of vehicles remains a challenging task, particularly when aiming to generate standardized models from multiple samples of the same make, model, body type, and year.
📊
Key Findings & Data:This study aims to overcome these limitations by presenting a novel automated pipeline for standardized 3D vehicle reconstruction.
💡
Conclusion & Impact:This contribution establishes a standardized benchmark resource for tasks such as novel view synthesis, damage assessment, and vehicle appearance modeling, representing one of the first publicly available datasets of vehicle-class-specific 3D Gaussian Splatting reconstructions derived from heterogeneous multi-source collections.
🔬Design:Modeling & Simulation
📍Scope:Comprehensive Sample
⚙️Variables:3B Gauss Saçılımı · 3B Araç Rekonstrüksiyonu · Hareketten Yapı Çıkarımı

Abstract

Background/Aim: Three-dimensional (3D) reconstruction of vehicles remains a challenging task, particularly when aiming to generate standardized models from multiple samples of the same make, model, body type, and year. While traditional Structure-from-Motion approaches excel at reconstructing individual objects, they struggle to create generalized models from diverse multi-view datasets exhibiting variations in lighting, background, and appearance. This study aims to overcome these limitations by presenting a novel automated pipeline for standardized 3D vehicle reconstruction. Methods: The proposed methodology employs a three-stage approach: (1) preprocessing with background removal and adaptive histogram equalization to isolate vehicle regions and normalize illumination; (2) deep learning-based orientation classification utilizing an EfficientNet-B0 architecture to categorize images into eight directional views; and (3) sparse reconstruction via hierarchical localization and COLMAP, implementing a neighbor-based image pairing strategy. Results: The developed pipeline was applied to a diverse collection of heterogeneous, multi-source vehicle images. This process resulted in the successful generation of 3D Gaussian Splatting models for over a hundred distinct vehicle classes. Conclusion: This study provides a robust framework for scalable 3D vehicle modeling. To support reproducibility and facilitate future research in generative AI and autonomous systems, the resulting dataset will be made publicly available. This contribution establishes a standardized benchmark resource for tasks such as novel view synthesis, damage assessment, and vehicle appearance modeling, representing one of the first publicly available datasets of vehicle-class-specific 3D Gaussian Splatting reconstructions derived from heterogeneous multi-source collections.

Özet

Arka Plan/Amaç: Araçların üç boyutlu (3B) rekonstrüksiyonu, özellikle aynı marka, model, kasa tipi ve üretim yılına ait çok sayıda örnekten standartlaştırılmış modeller üretilmesi amaçlandığında zorlu bir görev olmaya devam etmektedir. Geleneksel Hareketten Yapı Çıkarımı yaklaşımları bireysel nesneleri yeniden oluşturmada başarılı olsalar da; aydınlatma, arka plan ve görünümde farklılıklar sergileyen çeşitli çoklu-görünüm veri setlerinden genelleştirilmiş modeller oluşturmakta zorlanmaktadır. Bu çalışma, standartlaştırılmış 3B araç rekonstrüksiyonu için yeni ve otomatik bir boru hattı sunarak bu sınırlamaları ele almayı amaçlamaktadır. Yöntemler: Önerilen metodoloji üç aşamalı bir yaklaşım kullanmaktadır: (1) araç bölgelerini izole etmek ve aydınlatmayı normalleştirmek için arka planın kaldırılması ve uyarlanabilir histogram eşitlemesi ile ön işleme; (2) görüntüleri sekiz yönsel görünüme ayırmak için EfficientNet-B0 mimarisini kullanan derin öğrenme tabanlı yönelim sınıflandırması ve (3) komşuluk tabanlı bir görüntü eşleştirme stratejisi uygulayarak COLMAP ile hiyerarşik konumlandırma aracılığıyla seyrek rekonstrüksiyonu. Bulgular: Geliştirilen boru hattı, çok çeşitli, heterojen ve çok kaynaklı araç görüntüsü koleksiyonuna başarıyla uygulanmıştır. Bu süreç, yüzden fazla farklı araç sınıfı için 3B Gauss Saçılımı modellerinin üretilmesiyle sonuçlanmıştır. Sonuç: Bu çalışma, ölçeklenebilir 3B araç modellemesi için sağlam bir çerçeve sunmaktadır. Tekrarlanabilirliği desteklemek ve üretken yapay zeka ile otonom sistemlerdeki gelecekteki araştırmaları kolaylaştırmak amacıyla, elde edilen veri seti kamuya açık hale getirilecektir. Bu katkı; yeni görünüm sentezi, hasar tespiti ve araç görünümü modelleme gibi görevler için standartlaştırılmış bir kıyaslama kaynağı oluşturmakta olup, heterojen çok kaynaklı görüntü koleksiyonlarından elde edilen araca özgü 3B Gauss Saçılımı rekonstrüksiyonlarının ilk halka açık veri setlerinden birini temsil etmektedir.

🔖

Cite This Article

Uslu, M. A., Berkaya, S. K. (2026). A Multi-Stage Pipeline for 3D Reconstruction of Vehicles from Heterogeneous Multi-View Images. Erciyes Üniversitesi Fen Bilimleri Enstitüsü Fen Bilimleri Dergisi, Cilt: 42 Sayı: 2. https://doi.org/10.65520/erciyesfen.1918173

📄 Full Text PDF Viewer

📄

PDF görüntüleyici hazırlanıyor...

References (35)

  1. Verykokou, S., & Ioannidis, C. (2023). An overview on image-based and scanner-based 3D modeling technologies. Sensors, 23(2), 596.
  2. Liu, Y., Zhu, K., Wu, G., Ren, Y., Liu, B., Liu, Y., & Shan, J. (2023). MV-DeepSDF: Implicit modeling with multi-sweep point clouds for 3D vehicle reconstruction in autonomous driving. In Proceedings of the IEEE/CVF International Conference on Computer Vision (pp. 8306-8316).
  3. Mildenhall, B., Srinivasan, P. P., Tancik, M., Barron, J. T., Ramamoorthi, R., & Ng, R. (2021). Nerf: Representing scenes as neural radiance fields for view synthesis. Communications of the ACM, 65(1), 99–106.
  4. Kerbl, B., Kopanas, G., Leimkühler, T., & Drettakis, G. (2023). 3d gaussian splatting for real-time radiance field rendering. ACM Transactions on Graphics, 42(4), 139:1–14.
  5. Yang, L., Luo, P., Change Loy, C., & Tang, X. (2015). A large-scale car dataset for fine-grained categorization and verification. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 3973–3981).
  6. Hayee, S., Hussain, F., Yousaf, M. H., Yasir, M., Ahmad, S., & Viriri, S. (2025). A benchmark dataset and methodology for fine grained vehicle make and model classification. Frontiers in Computer Science, 7, 1561899.
  7. Tan, M., & Le, Q. (2019, May). EfficientNet: Rethinking model scaling for convolutional neural networks. In International conference on machine learning (pp. 6105-6114). PMLR
  8. Ruan, G., Hu, T., Ding, C., Yang, K., Kong, F., Cheng, J., & Yan, R. (2025). Fine-grained vehicle recognition under low light conditions using EfficientNet and image enhancement on LiDAR point cloud data. Scientific reports, 15(1), 4691.
  9. Asmitha, U., Roshan Tushar, S., Sowmya, V., & Soman, K. P. (2022, September). Ensemble deep learning models for vehicle classification in motorized traffic analysis. In International Conference on Innovative Computing and Communications: Proceedings of ICICC 2022, Volume 2 (pp. 185-192). Singapore: Springer Nature Singapore.
  10. Kim, J. (2024). Deep learning-based vehicle type and color classification to support safe autonomous driving. Applied Sciences, 14(4), 1600.
  11. Zhu, H., Zhang, Z., Zhao, J., Duan, H., Ding, Y., Xiao, X., & Yuan, J. (2025). Scene reconstruction techniques for autonomous driving: a review of 3D Gaussian splatting. Artificial Intelligence Review, 58(1).
  12. Zhou, X., Lin, Z., Shan, X., Wang, Y., Sun, D., & Yang, M. H. (2024). Drivinggaussian: Composite gaussian splatting for surrounding dynamic autonomous driving scenes. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 21634-21643).
  13. Yan, Y., Lin, H., Zhou, C., Wang, W., Sun, H., Zhan, K.,. & Peng, S. (2024, September). Street gaussians: Modeling dynamic urban scenes with gaussian splatting. In European Conference on Computer Vision (pp. 156-173). Cham: Springer Nature Switzerland.
  14. Tian, Q., Tan, X., Xie, Y., & Ma, L. (2025, April). Drivingforward: Feed-forward 3d gaussian splatting for driving scene reconstruction from flexible surround-view input. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 39, No. 7, pp. 7374-7382).
  15. Du, X., Wang, Y., Sun, H., Wu, Z., Sheng, H., Wang, S.,. & Yu, X. (2025). 3drealcar: An in-the-wild rgb-d car dataset with 360-degree views. In Proceedings of the IEEE/CVF International Conference on Computer Vision (pp. 26488-26498).
  16. Gatis, D. (2025, February 21). rembg (Version 2.0.66) [Computer software]. GitHub. https://github.com/danielgatis/rembg
  17. Zuiderveld, K. (1994). Contrast limited adaptive histogram equalization. In P. S. Heckbert (Ed.), Graphics gems IV (pp. 474–485). Academic Press.
  18. He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 770-778).
  19. Kornblith, S., Shlens, J., & Le, Q. V. (2019). Do better imagenet models transfer better? In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 2661-2671).
  20. Schönberger, J. L., & Frahm, J. M. (2016). Structure-from-motion revisited. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 4104-4113).
  21. Furukawa, Y., & Hernández, C. (2015). Multi-view stereo: A tutorial. Foundations and trends® in Computer Graphics and Vision, 9(1-2), 1-148.
  22. Sarlin, P. E., Cadena, C., Siegwart, R., & Dymczyk, M. (2019). From coarse to fine: Robust hierarchical localization at large scale. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 12716-12725).
  23. DeTone, D., Malisiewicz, T., & Rabinovich, A. (2018). Superpoint: Self-supervised interest point detection and description. In Proceedings of the IEEE conference on computer vision and pattern recognition workshops (pp. 224-236).
  24. Sarlin, P. E., DeTone, D., Malisiewicz, T., & Rabinovich, A. (2020). Superglue: Learning feature matching with graph neural networks. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 4938-4947).
  25. Martin-Brualla, R., Radwan, N., Sajjadi, M. S., Barron, J. T., Dosovitskiy, A., & Duckworth, D. (2021). Nerf in the wild: Neural radiance fields for unconstrained photo collections. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 7210-7219).
  26. Kulhanek, J., Peng, S., Kukelova, Z., Pollefeys, M., & Sattler, T. (2024). Wildgaussians: 3d gaussian splatting in the wild. arXiv preprint arXiv:2407.08447.
  27. Ye, V., Li, R., Kerr, J., Turkulainen, M., Yi, B., Pan, Z., Seiskari, O., Ye, J., Hu, J., Tancik, M. & Angjoo K. (2025). gsplat: An open-source library for Gaussian splatting. Journal of Machine Learning Research, 26(34), 1-17.
  28. Wang, Z., Bovik, A. C., Sheikh, H. R., & Simoncelli, E. P. (2004). Image quality assessment: from error visibility to structural similarity. IEEE transactions on image processing, 13(4), 600-612.
  29. Zhang, R., Isola, P., Efros, A. A., Shechtman, E., & Wang, O. (2018). The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 586-595).
  30. Tyszkiewicz, M., Fua, P., & Trulls, E. (2020). Disk: Learning local features with policy gradient. Advances in neural information processing systems, 33, 14254-14265.
  31. Lindenberger, P., Sarlin, P. E., & Pollefeys, M. (2023). Lightglue: Local feature matching at light speed. In Proceedings of the IEEE/CVF international conference on computer vision (pp. 17627-17638).
  32. Zhao, X., Wu, X., Chen, W., Chen, P. C., Xu, Q., & Li, Z. (2023). Aliked: A lighter keypoint and descriptor extraction network via deformable transformation. IEEE Transactions on Instrumentation and Measurement, 72, 1-16.
  33. Lowe, D. G. (2004). Distinctive image features from scale-invariant keypoints. International journal of computer vision, 60(2), 91-110.
  34. Yan, Y., Lin, H., Zhou, C., Wang, W., Sun, H., Zhan, K.,. & Peng, S. (2024, September). Street gaussians: Modeling dynamic urban scenes with gaussian splatting. In European Conference on Computer Vision (pp. 156-173). Cham: Springer Nature Switzerland.
  35. Du, X., Wang, Y., Sun, H., Wu, Z., Sheng, H., Wang, S.,. & Yu, X. (2025). 3drealcar: An in-the-wild rgb-d car dataset with 360-degree views. In Proceedings of the IEEE/CVF International Conference on Computer Vision (pp. 26488-26498).