CUDA-Based Parallel Intra Prediction and Transform Architecture for Hevc Intra-Frame Image Coding

Öz

Computations related to intra prediction and transform in High Efficiency Video Coding are highly computational demanding, especially when multiple intra prediction modes have to be computed for block-based image encoding. In this paper, we propose a parallel CUDA implementation of HEVC intra frame prediction and integer transform operations on an embedded GPU platform. The presented implementation parallelizes HEVC intra prediction modes (35 in total, divided into planar, DC, and angular modes) and implements an efficient transform computation utilizing shared memory allocation. The proposed implementation was tested on NVIDIA Jetson Xavier NX platform using standardized test images and various quantization parameter settings. The results demonstrate that our solution accelerates the computations by up to 105.17× in terms of kernel computations compared to the sequential CPU implementation. Taking into account CPU–GPU memory transfer and synchronization overheads, the acceleration factor for the end-to-end process is 1.62×. In terms of the reconstructed image quality, there is a small decrease in the average PSNR value equal to about 0.65 dB. It is mainly caused by implementation-related differences in parallel computation, rounding, and quantization procedures. These results show that intra-frame prediction and transform stages in HEVC video codec can be efficiently parallelized using CUDA; however, there are still memory transfer and synchronization overheads limiting end-to-end performance. Our implementation can serve as a building block of a fast HEVC encoder but not as a full-fledged video encoder itself. Source code and dataset for this research are freely available on Zenodo with the DOI: https://doi.org/10.5281/zenodo.17286247.

Anahtar Kelimeler

• HEVC intra prediction
• CUDA
• Integer DCT
• Embedded GPU
• Intra-frame image coding
• Block-based image compression

Etik Beyan

Ethics committee approval was not required for this study because of there was no study on animals or humans.

Kaynakça

1. Akman, A., & Cekli, S. (2023). Efficient 2D DCT architecture based on approximate compressors for image compression with HEVC intra-prediction. Journal of Real-Time Image Processing, 20(22). https://doi.org/10.1007/s11554-023-01261-3
2. Bloch, A., Casale-Brunet, S., & Mattavelli, M. (2023). Design space exploration for partitioning dataflow programs on CPU + GPU heterogeneous systems. Journal of Signal Processing Systems, 95. 1219–1229. https://doi.org/10.1007/s11265-023-01884-6
3. Bross, B., Han, W.-J., Sullivian, G. J., Ohm, J.-R., & Wiegand, T. (2023). High Efficiency Video Coding (HEVC) text specification draft 9. http://phenix.itsudparis.eu/jct/doc_end_user/current_document.php?id=6803 (accessed on 15 June 2023).
4. Ghorpade, J., Parande, J., Kulkarni, M., & Bawaskar, A. (2012). GPGPU processing in CUDA architecture, arXiv preprint arXiv:1202.4347.
5. Harris, M. (2019). How to optimize data transfers in CUDA C/C++. NVIDIA Developer Blog. https://devblogs.nvidia.com/how-optimize-data-transfers-cuda-cc/
6. Imen, W., Amna. M., Fatma. B., Ezahra. S. F., & Masmoudi. N. (2022). Fast HEVC intra-CU decision partition algorithm with modified LeNet-5 and AlexNet. Signal. Image and Video Processing. 16(7). 1811–1819. https://doi.org/10.1007/s11760-022-02139-w
7. James, R., Abo-Zahhad. M., Inoue, K., & Sayed, M. S. (2024). Performance evaluation of all intra Kvazaar and x265 HEVC encoders on embedded system NVIDIA Jetson platform. Journal of Real-Time Image Processing. 21(67). https://doi.org/10.1007/s11554-024-01429-5
8. Jingjing, Y., & Kehua, J. (2019). A fast intra prediction algorithm with simplified prediction modes based on utilization rates. 2019 IEEE/ACIS 18th International Conference on Computer and Information Science (ICIS). 225–229. Beijing. China.

9. Juncheng, M., Falei, L., Shanshe, W., Nan. Z., & Siwei. M. (2015). Parallel intra coding for HEVC on CPU plus GPU platform. 2015 Visual Communications and Image Processing (VCIP), 1–4.
10. Kaplan, M., & Akman, A. (2024). Parallel implementation of Discrete Cosine Transform (DCT) methods on GPU for HEVC. In M. N. Seyman (Ed.). 2nd International Congress of Electrical and Computer Engineering: ICECENG 2023. EAI/Springer Innovations in Communication and Computing, Springer. https://doi.org/10.1007/978-3-031-52760-9_20
11. Khemiri, R., Kibeya, H., Sayadi, F. E., Bahri, N., Atri, M., & Masmoudi, N. (2018). Optimization of HEVC motion estimation exploiting SAD and SSD GPU-based implementation, IET Image Processing, 12(2), 243–253. https://doi.org/10.1049/iet-ipr.2017.0474
12. Li, W., Ou, W., Wang, H., Wu, Y., & Zhong, Y. (2025). Jointly probability-based fast intra prediction algorithm for spatial SHVC. EURASIP Journal on Image and Video Processing, 2025, 20. https://doi.org/10.1186/s13640-025-00683-2
13. Lini X., Sun, S., Huang, W., Sheng, B., Li, P., & Feng, D. D. (2023). EAPT: Efficient attention pyramid transformer for image processing. IEEE Transactions on Multimedia, 25, 50–61. https://doi.org/10.1109/TMM.2021.3120873
14. Lopez-Granado, O., Migallon, H., Alcocer Espinosa, E., Gutierrez, R., Van Wallendael, G., & Malumbres, M. P. (2025). Performance, limitations, and design issues of the integration of a hardware-based IME module with HEVC video encoder software. IEEE Access, 13, 113390–113406. https://doi.org/10.1109/ACCESS.2025.3581961
15. Mate, Č., Alen, D., Leon, D., Igor, P., & Mario, K. (2020). Performance engineering for HEVC transform and quantization kernel on GPUs. Automatika, 61(3), 325–333. https://doi.org/10.1080/00051144.2020.1752046
16. NVIDIA. (2019). CUDA refresher: The CUDA programming model, NVIDIA Developer Blog. https://developer.nvidia.com/blog/cuda-refresher-cuda-programming-model/ (accessed on 18 July 2023).
17. NVIDIA. (2023). NVIDIA Jetson Xavier NX developer kit. https://openzeka.com/wpcontent/uploads/2020/03/Jetson_Xavier_NX_Developer_Kit-One-Pager.pdf (accessed on 23 July 2023).
18. Pietri, N., Nocent, O., Robert, P., & Schmutz, A. (2025). Real-time live compression of dynamic 3D meshes for client-side GPU cloud gaming using skinning decomposition, The Visual Computer, Advance online publication. https://doi.org/10.1007/s00371-025-04112-3
19. Radicke, S., Hahn, J., Wang, Q., & Grecos, C. (2016). A parallel HEVC intra prediction algorithm for heterogeneous CPU+GPU platforms. IEEE Transactions on Broadcasting. 62(1), 103–119.
20. Rathee, N., Sharma, T., Yadav, J., et al. (2026). High-performance FPGA implementation of HEVC with dynamic programming for optimized intra-frame prediction. Scientific Reports, 16, 21241. https://doi.org/10.1038/s41598-026-51672-y
21. Ryoo, S., Rodrigues, C., Baghsorkhi, S, S., et al. (2008). Optimization principles and application performance evaluation of a multithreaded GPU using CUDA. Proceedings of the 13th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming (PPoPP), 73–82. ACM.
22. Singhadia, A., Bante, P., & Chakrabarti, I. (2019). A novel algorithmic approach for efficient realization of 2-D-DCT architecture for HEVC, IEEE Transactions on Consumer Electronics, 65(3), 264–273.
23. Tariq, J. (2020). High-performance intra-mode accelerator for HEVC, The Visual Computer, 36(8), 1603–1617. https://doi.org/10.1007/s00371-019-01764-w
24. Thurley, M, J., & Danell, V. (2012). Fast morphological image processing open-source extensions for GPU processing with CUDA, IEEE Journal of Selected Topics in Signal Processing, 849–855.
25. Wang, S., Liu, J., Zhang, H., & Hu, Y. (2025). A fast algorithm of HEVC intra mode decision. Journal of Shanghai University (Natural Science Edition), 31(3), 561–570. https://doi.org/10.12066/j.issn.1007-2861.2639
26. Zhu, J., Zhang, Q., Fei, L., Cai, R., Xie, Y., Sheng, B., & Yang, X. (2023). FFFN: Frame-by-frame feedback fusion network for video super-resolution, IEEE Transactions on Multimedia, 25, 6821–6834. https://doi.org/10.1109/TMM.2022.3214776