Journal of Information Processing Systems

Korea Information Processing Society (한국정보처리학회)
권호 표지
DOI QR코드

DOI QR Code

Resource Efficient AI Service Framework Associated with a Real-Time Object Detector

  • Jun-Hyuk Choi (Dept. of Embedded Systems Engineering, Incheon National University) ;
  • Jeonghun Lee (Dept. of Embedded Systems Engineering, Incheon National University) ;
  • Kwang-il Hwang (Dept. of Embedded Systems Engineering, Incheon National University)
  • Received : 2023.02.01
  • Accepted : 2023.03.02
  • Published : 2023.08.31
Download PDF
⟨ Previous Next ⟩

Abstract

This paper deals with a resource efficient artificial intelligence (AI) service architecture for multi-channel video streams. As an AI service, we consider the object detection model, which is the most representative for video applications. Since most object detection models are basically designed for a single channel video stream, the utilization of the additional resource for multi-channel video stream processing is inevitable. Therefore, we propose a resource efficient AI service framework, which can be associated with various AI service models. Our framework is designed based on the modular architecture, which consists of adaptive frame control (AFC) Manager, multiplexer (MUX), adaptive channel selector (ACS), and YOLO interface units. In order to run only a single YOLO process without regard to the number of channels, we propose a novel approach efficiently dealing with multi-channel input streams. Through the experiment, it is shown that the framework is capable of performing object detection service with minimum resource utilization even in the circumstance of multi-channel streams. In addition, each service can be guaranteed within a deadline.

Keywords

Acknowledgement

This work was supported by Incheon National University Research Grant 2019.

References

  1. A. Bochkovskiy, C. Y. Wang, and H. Y. M. Liao, "Yolov4: optimal speed and accuracy of object detection," 2020 [Online]. Available: https://arxiv.org/abs/2004.10934.
  2. J. Parsola, D. Gangodkar, and A. Mittal, "Post event investigation of multi-stream video data utilizing Hadoop cluster," International Journal of Electrical & Computer Engineering, vol. 8, no. 6, pp. 5089-5097, 2018. https://doi.org/10.11591/ijece.v8i6.pp5089-5097
  3. S. Karimi-Mansoub, R. Abri, and A. Yarici, "Concurrent real-time object detection on multiple live streams using optimization CPU and GPU resources in YOLOv3," in Proceedings of the 4th International Conference on Advances in Signal, Image and Video Processing, Athens, Greece, 2019, pp. 23-28.
  4. L. Tan, X. Dong, Y. Ma, and C. Yu, "A multiple object tracking algorithm based on YOLO detection," in Proceedings of 2018 11th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI), Beijing, China, 2018, pp. 1-5. https://doi.org/10.1109/CISP-BMEI.2018.8633009
  5. D. Kim and S. Park, "An intelligent collaboration framework between edge camera and video analysis system," in Proceedings of 2018 International Conference on Electronics, Information, and Communication (ICEIC), Honolulu, HI, 2018, pp. 1-3. https://doi.org/10.23919/ELINFOCOM.2018.8330653
  6. G. Mattela, C. Pal, M. Tripathi, R. Gavval, and A. Acharyya, "Enterprise class deep neural network architecture for recognizing objects and faces for surveillance systems," in Proceedings of 2019 11th International Conference on Communication Systems & Networks (COMSNETS), Bengaluru, India, 2019, pp. 607-612. https://doi.org/10.1109/COMSNETS.2019.8711399
  7. S. Aslan and S. C. Ileri, "Performance analysis of ARM big.LITTLE architecture based mobile processor with multi-thread face detection," in Proceedings of 2019 4th International Conference on Computer Science and Engineering (UBMK), Samsun, Turkey, 2019, pp. 336-339. https://doi.org/10.1109/UBMK.2019.8907058
  8. X. Wang, "An efficient end-to-end object detection pipeline on GPU using CUDA," master's thesis, Eindhoven University of Technology, Eindhoven, The Netherlands, 2019.
  9. S. Liu, G. Liu, and H. Zhou, "A robust parallel object tracking method for illumination variations," Mobile Networks and Applications, vol. 24, pp. 5-17, 2019. https://doi.org/10.1007/s11036-018-1134-8
  10. J. Lee and K. I. Hwang, "YOLO with adaptive frame control for real-time object detection applications," Multimedia Tools and Applications, vol. 81, pp. 36375-36396, 2022. https://doi.org/10.1007/s11042-021-11480-0