Improving Image Analysis and Processing Performance on the RISC-V Platform with Lichee Pi 4A
https://doi.org/10.15514/ISPRAS-2025-37(5)-12
Abstract
The study explores optimization methods for improving image processing performance on the RISC‑V platform with Lichee Pi 4A. The research focuses on real-time video processing within a microservice-based self-service system. Several existing optimization strategies are considered and evaluated, including neural network model optimization, hardware acceleration using RVV vector instructions and leveraging the built-in Neural Processing Unit (NPU). The profiling results on existing strategies indicate that object detection and feature extraction consume the most computation resources. In order to eliminate the performance gap, the model quantization to INT8 format is implemented, that allows to reduce memory usage and inference latency. Additionally, a modified ONNX Runtime version is deployed to support NPU acceleration. These improvements led to 75% reduction in model size and a 35% decrease in inference latency. The study concludes that hardware-aware optimizations significantly enchase performance on the RISC-V (Lichee Pi 4A) platform. The main issue encountered is the low processing speed on Lichee Pi 4A, with a current frame rate of only 0.05 FPS, which in unsuitable for practical usage.
About the Authors
Nikita Ivanovich CHEREPANOV
Peter the Great St. Petersburg Polytechnic University
Russian Federation
Russian Federation
A master's student at the Higher School of Software Engineering at Peter the Great St. Petersburg Polytechnic University. In 2025, he got bachelor degree by graduating from Peter the Great St. Petersburg Polytechnic University with a specialty in “Technology for developing and maintaining a high-quality software product”. Research interests: software architectures, RISC-V, machine learning, computer vision, artificial intelligence.
Nadegda Olegovna STEPINA
Peter the Great St. Petersburg Polytechnic University
Russian Federation
Russian Federation
An assistant at the Higher School of Software Engineering at Peter the Great St. Petersburg Polytechnic University. In 2023, she graduated from the St. Petersburg State Polytechnic University with a degree in Software Engineering. In 2024, she became a postgraduate student in the field of Mathematical and Software Support for Computing Machines, Complexes, and Computer Networks. Her research interests include software development, machine learning, high-performance computing, IoT, and embedded systems.
Igor Valerievich NIKIFOROV
Peter the Great St. Petersburg Polytechnic University
Russian Federation
Russian Federation
In 2011, he graduated from St. Petersburg State Polytechnic University with a degree in «Computer Science and Automated Systems Software». He got his Cand. Sci. (Tech.) degree in Mathematical and software support for computers, complexes and computer networks in 2014. He is an Associate Professor at the Higher School of Software Engineering at Peter the Great St. Petersburg Polytechnic University. He is the author of more than 100 scientific publications. Research interests – software engineering, simulation modeling, big data analytics, distributed computing.
References
1. Cui E., Li T. Wei Q. RISC-V instruction set architecture extensions. A survey. IEEE Access 11, 2023, 24696–24711. DOI: 10.1109/ACCESS.2023.3246491.
2. Shen Y. Computer Vision: Technologies and Applications. Applied and Computational Engineering, vol. 163, no. 1, pp. 35–41, Jun. 2025, DOI: 10.54254/2755-2721/2025.23817.
3. Ali W. Exploring Instruction Set Architectural Variations: x86, ARM, and RISC-V in Compute-Intensive Applications, Aug. 2023, DOI: 10.36227/techrxiv.24026736.
4. Han C., Chang C., Srivastava S., Lu Y. Scalable Complex Event Processing on Video Streams. Proc. ACM on Management of Data, vol. 3, no. 3, pp. 1–29, Jun. 2025, DOI: 10.1145/3725419.
5. Borysenko V., Borysenko T. Modern approaches of design software applications based on microservice architecture in computer and information systems and technologies. Apr. 2020, DOI: 10.30837/IVcsitic2020201441.
6. Bhatnagar S., Mahant R. Designing Microservices in The Art of Decoding Microservices: An In-Depth Exploration of Modern Software Architecture. Launch IT, 2025, pp. 135-192.
7. Domenech-Asensi G., Garrigos J., Lopez P., Brea V., Cabello D. Real time architectures for the Scale Invariant Feature Transform algorithm. CNNA 2016; 15th International Workshop on Cellular Nanoscale Networks and their Applications. Dresden, Germany, 2016, pp. 1-2.
8. Obukhov A., Dedov D., Volkov A., Rybachok M. Technology for Improving the Accuracy of Predicting the Position and Speed of Human Movement Based on Machine Learning Models. Technologies, vol. 13, no. 3, p. 101, Mar. 2025, DOI: 10.3390/technologies13030101.
9. Qin X., Liu X., Han J. A CNN Hardware Accelerator Designed for YOLO Algorithm Based on RISC-V SoC. Proc. IEEE Int. Conf. ASIC, Kunming, China, 2021, pp. 1-4, DOI: 10.1109/ASICON52560.2021.9620500.
10. Srivastava S. K, Srivastava A. K., Allam S., Lilaramani D. Comparative analysis on Deep Convolution Neural Network models using Pytorch and OpenCV DNN frameworks for identifying optimum fruit detection solution on RISC-V architecture. IEEE Mysore Sub Section International Conference (MysuruCon), Hassan, India, 2021, pp. 738-743, DOI: 10.1109/MysuruCon52639.2021.9641594.
11. Chen Y.-R. Experiments and optimizations for TVM on RISC-V Architectures with P Extension. International Symposium on VLSI Design, Automation and Test (VLSI-DAT), Hsinchu, Taiwan, 2020, pp. 1-4, DOI: 10.1109/VLSI-DAT49148.2020.9196477.
12. Yu M.-S., Chang H.-C., Wang C.-T., Tien Y.-W. Optimizing computer vision algorithms with TVM on VLIW architecture based on RVV. The Journal of Supercomputing, vol. 81, no. 1, Nov. 2024, DOI: 10.1007/s11227-024-06530-x.
13. Jajal P, Jiang W, Tewari A, Kocinare E, Woo J, Sarraf A. Interoperability in deep learning: a user survey and failure analysis of ONNX model converters. In: Proc. 33rd ACM SIGSOFT International Symposium on Software Testing and Analysis. New York: ACM; 2024. p. 1466–1478, DOI: 10.1145/3650212.3680374.
14. Fusaomi N., Shingo S., Ryoma A., Keigo W., Maki K. H. Evaluation of Interoperability of CNN Models between MATLAB and Python Environments Using ONNX Runtime Model. AI, Computer Science and Robotics Technology 3(1), 1–13. 2024, DOI: 10.5772/acrt.20240043.
15. Sohan M., Ram T. S., Ch V. R. R. A Review on YOLOv8 and Its Advancements. Data Intelligence and Cognitive Informatics, Jan. 2024, pp. 529–545, DOI: 10.1007/978-981-99-7962-2_39.
16. Almeyda S., Davila A.: Process Improvement in Software Requirements Engineering: A Systematic Mapping Study. Programming and Computer Software, 48, Aug. 2022, pp. 513–533. DOI: 10.1134/S0361768822080084.
17. Lunev D., Poletykin S., Kudryavtsev D. Brain-computer interfaces: Technology overview and modern solutions. Modern Innovations, Systems and Technologies, vol. 2, no. 3, Jul. 2022, pp. 0117-0126, DOI: 10.47813/2782-2818-2022-2-3-01170126.
18. Tsekhmystro R., Rubel O., Prysiazhniuk O., Lukin V. V. Impact of distortions in UAV images on quality and accuracy of object localization. radioelectronic and computer systems, Jan. 2025, DOI: 10.32620/reks.2024.4.05.
19. Fischer L., Wollstadt P. Precision and Recall Reject Curves for Classification. Aug. 2023, DOI: 10.48550/arXiv.2308.08381.
20. Boyd K., Eng K. H., Page C. D. Area under the Precision-Recall Curve: Point Estimates and Confidence Intervals. Joint European Conference on Machine Learning and Knowledge Discovery in Databases, Lecture Notes in Computer Science, vol. 8190, pp. 451–466, Sep. 2013, DOI: 10.1007/978-3-642-40994-3_29.
21. Sinha E., Kumar A., Tyagi A. OpenCV for Computer Vision Applications. International Journal For Multidisciplinary Research, vol. 7, no. 3, May 2025, DOI: 10.36948/ijfmr.2025.v07i03.44280.
22. Chinnaraju A. Benchmarking cross‑platform AI: Web Assembly, ONNX Runtime and TVM for Real‑Time Web, Mobile, and IoT Deployment. World Journal of Advanced Research and Reviews, vol. 26, no. 2, pp. 1937–1963, May 2025, DOI: 10.30574/wjarr.2025.26.2.1832.
23. Yang S., Lu T. T3 SOC design flow case study: Design a video processing pipeline. ASIC, ASICON '07. 7th International Conference, Nov. 2007, DOI: 10.1109/ICASIC.2007.4415551.
24. Jindal K. Design and Implementation of an Embedded Image Processing System on Zynq ZedBoard: A VLSI Perspective. International Journal for Research in Applied Science and Engineering Technology, vol. 13, no. 5, pp. 5141–5145, May 2025, DOI: 10.22214/ijraset.2025.71372.
25. Smirnov E., Timoshenko D., Andrianov S. Comparison of Regularization Methods for ImageNet Classification with Deep Convolutional Neural Networks. AASRI Procedia, vol. 6, pp. 89–94, Dec. 2014, DOI: 10.1016/j.aasri.2014.05.013.
26. Pujari S. D., Pawar M. M., Wadekar M. Multi-Classification of Breast Histopathological Image Using Xception: Deep Learning with Depthwise Separable Convolutions Model. Techno-Societal, pp. 539–546, May 2021, DOI: 10.1007/978-3-030-69921-5_54.
27. Wang S., Wang X., Xu Z., Chen B. Optimizing CNN Computation Using RISC-V Custom Instruction Sets for Edge Platforms. IEEE Trans. Comput, May 2024, pp. 1-14, DOI: 10.1109/TC.2024.3362060.
28. Titopoulos V., Alexakis G., Nicopoulos C., Dimitrakopoulos G. Efficient Implementation of RISC-V Vector Permutation Instructions. arXiv:2505.07112, May 2025, DOI: 10.48550/arXiv.2505.07112.
29. Yuan T., Liu W., Han J., Lombardi F. High Performance CNN Accelerators Based on Hardware and Algorithm Co-Optimization. IEEE Trans. Circuits Syst. I, Reg. Papers, Oct. 2020, pp. 1-14, DOI: 10.1109/TCSI.2020.3030663.
30. Jin S., Qi S., Dai Y., Hu Y. Design of Convolutional Neural Network Accelerator Based on RISC-V. Proc. 10th Int. Conf. Appl. Tech. Cyber Intell. (ICATCI 2022), 2023, pp. 446-454. DOI: 10.1007/978-3-031-29097-8_53.
31. Cono D’Elia D., Demetrescu C. Ball-Larus Path Profiling across Multiple Loop Iterations. SIGPLAN Not. 48, 10 (oct 2013), pp. 373-390, DOI :10.1145/2544173.2509521.
32. Agarwal R., Deshmukh R., Borhade P., Murarka S. Image Classification using Parallel CPU and GPU Computing. Int. J. Eng. Adv. Technol., vol. 9, no. 4, Apr. 2020, pp. 5, DOI: 10.35940/ijeat.D7870.049420.
33. Shanthi M., Anthony Irudhayaraj A. Multithreading - An Efficient Technique for Enhancing Application Performance. International Journal of Recent Trends in Engineering, Vol 2, No. 4, Nov. 2009, pp. 165-167, DOI: 10.22146/ijccs.57594.
Review
For citations:
CHEREPANOV N.I., STEPINA N.O., NIKIFOROV I.V. Improving Image Analysis and Processing Performance on the RISC-V Platform with Lichee Pi 4A. Proceedings of the Institute for System Programming of the RAS (Proceedings of ISP RAS). 2025;37(5):157-172. https://doi.org/10.15514/ISPRAS-2025-37(5)-12
Email this article 