Abstract
The High-Efficiency Video Coding (HEVC) standard has emerged to achieve high coding efficiency performance while introducing several novel tools. One contributor key to the performance gain over Advanced Video Coding (AVC) H.264 is the unit partition structure that extended a large number of coding unit shapes (CU) ranging from 64\(\times \)64 to 8x8 to replace the macroblock structure of H.264/AVC. This pioneering module achieves a significant gap of compression performance at the expense of additional encoding complexity, which increases under HEVC All-intra configuration due to the use of the Rate-Distortion Optimization (RDO) process. Since I-frame significantly affects the coding efficiency, the main goal of the proposed work is to implement Convolutional Neural Network-based approaches to substitute the brute force RDO search without affecting the compression efficiency performance. LeNet-5-based and AlexNet-based approaches are developed to eliminate extensive computational time used to check all block decision candidates, which deeply optimize the HEVC coding unit partition module for All-intra configuration. In the first step, a database was created for the HEVC intra-mode to learn different models. Subsequently, modified LeNet-5 (M-LeNet-5) and modified AlexNet (M-AlexNet) models are implemented to predict the HEVC CU partition and their performances are compared. Experimental results indicated that the proposed algorithms could speed up the CU partition structure by reducing the intra-mode encoding time up to 85% and 75% with M-LeNet-5 and M-Alex-Net, respectively.
Similar content being viewed by others
References
Gary, S., Jens-Rainer, J.O., Woo-Jin, H., Wiegand, T.: Overview of the high efficiency video coding (HEVC) standard. IEEE Trans. Circuits Syst. Video Technol. 22(12), 1649–1668 (2012)
Thomas, W., Gary, S., Gisle, J.B., Luthra, A.: Overview of the H 264/AVC video coding standard. IEEE Trans. Circuits Syst. Video Technol. 13(7), 560–576 (2003)
Seunghyun, C., Munchurl, K.: Fast CU splitting and pruning for suboptimal CU partitioning in HEVC intra coding. IEEE Trans. Circuits Syst. Video Technol. 23(9), 1555–1564 (2013)
Mallikarachchi, T., Fernando, A., Arachchi, H.K.: Efficient coding unit size selection based on texture analysis for HEVC intra prediction. In: 2014 IEEE International Conference on Multimedia and Expo (ICME), pp. 1–6 (2014)
Yun, Z., Na, L., Sam, K., Zeng, H.: Statistical early termination and early skip models for fast mode decision in HEVC INTRA coding. ACM Trans. Multimedia Comput. Commun. Appl. 15(3), 1–23 (2019)
Fang, H., Chen, H., Chang, T.: Fast intra prediction algorithm and design for High Efficiency Video Coding. In: 2016 IEEE International Symposium on Circuits and Systems (ISCAS), pp. 1770–1773 (2016)
Kim, N., Jeon, S., Shim, H. J., Jeon, B., Lim, S., Ko, H.: Adaptive keypoint-based CU depth decision for HEVC intra coding. In: 2016 IEEE International Symposium on Broadband Multimedia Systems and Broadcasting (BMSB), pp. 1–3 (2016)
Yin, J., Yang, X., Lin, J., Chen, Y., Fang, R.: A fast block partitioning algorithm based on SVM for HEVC intra coding. In: Proceedings of the 2018 the 2nd International Conference on Video and Image Processing(ICVIP), pp. 29–31 (2018)
LIU, X., LI, Y., LIU, D., et al.: An adaptive CU size decision algorithm for HEVC intra prediction based on complexity classification using machine learning. IEEE Trans. Circuits Syst. Video Technol. 29(1), 144–155 (2017)
Du, B., Siu, W.-C., Yang, X. et al.: Fast CU partition strategy for HEVC intra-frame coding using learning approach via random forests. In: Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA). IEEE 2015, pp. 1085–1090 (2015)
Mai, X., Tianyi, L., Zulin, W., et al.: Reducing complexity of HEVC: a deep learning approach. IEEE Trans. Image Process. 27(10), 5044–5059 (2018)
Jun, S., Changsheng, G., Zhibo, C.: Asymmetric-Kernel CNN based fast CTU partition for HEVC intra coding. In: IEEE International Symposium on Circuits and Systems (ISCAS). IEEE 2019, pp. 1–5 (2019)
Guanwen, Z., Liyuan, X., Xiaocong L., et al.: A CNN-based coding unit partition in HEVC for video processing. In: IEEE International Conference on Real-time Computing and Robotics (RCAR). IEEE, pp. 273–276 (2019)
Hai-Che, T., Hung-Luen, F., Jia-Shung W., et al.: Complexity reduction on HEVC intra mode decision with modified LeNet-5. In: 2019 IEEE International Conference on Artificial Intelligence Circuits and Systems (AICAS). IEEE, pp. 20–24 (2019)
Zixi, W., Fan, L.: Convolutional neural network based low complexity HEVC intra encoder. Multimedia Tools Appl. 80(2), 2441–2460 (2021)
Yann, L., et al.: LeNet-5, convolutional neural networks. Science 20(5), 14 (2015)
Yann, L.: The MNIST database of handwritten digits. http://yann.lecun.com/exdb/mnist/ (1998)
Alex, K., Ilya, S., Geoffrey, H.E., et al.: Imagenet classification with deep convolutional neural networks. Adv Neural Inf. Process. Syst. 25, 1097–1105 (2012)
Frank, B., et al.: Common test conditions and software reference configurations. In: JCTVC-L1100, vol. 12, p. 7 (2013)
Xiph.org.: Xiph.org Video Test Media (2017). https://media.xiph.org/video/derf
Ultra VideoGroup : http://ultravideo.cs.tut.fi/#$testsequences
SJTU 4K Video Sequences:http://medialab.sjtu.edu.cn/web4k/index.html
JCT-VC, HM software,https://hevc.hhi.fraunhofer.de/svn/svn_HEVCSoftware/tags/HM-16.2/
Nikhil, K.: Introduction to pytorch. In: Deep learning with python. Apress, Berkeley, CA, pp. 195–208 (2017)
Bjontegaard, G.: Calculation of average PSNR difference between RD-curves. In: ITU-T Q.6/SG16 VCEG 13th Meeting, DocumentVCEG-M33 (2001)
JCT-VC: HM software,https://hevc.hhi.fraunhofer.de/svn/svn_HEVCSoftware/tags/HM-16.16/
Author information
Authors and Affiliations
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Imen, W., Amna, M., Fatma, B. et al. Fast HEVC intra-CU decision partition algorithm with modified LeNet-5 and AlexNet. SIViP 16, 1811–1819 (2022). https://doi.org/10.1007/s11760-022-02139-w
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11760-022-02139-w