Optimizing Real-Time Object Detection on Edge Devices: A Transfer Learning Approach

Authors

  • Harshad Lokhande, Sanjay Ganorkar

Keywords:

K-Means, DNN , IoT, Video Surveillance, edge computing, tinyML

Abstract

The issue of object detection in remote surveillance using edge devices presents a complex scenario, largely as a result of the limitations inherent in edge computing settings and the requirements for instantaneous data processing. Current video surveillance systems exhibit proficient video capture functionalities; however, data analysis at the server level is impeded by constraints in transmission power and the availability of cloud computing resources. Consequently, Internet of Things (IoT) devices are primarily relegated to the role of data acquisition. Our study proposes a novel fusion of ResNet18, K-Means clustering, and int8 quantization over tinyML, compressing the model to <100KB and enabling sub-1mW consumption on edge devices. This methodology extends the viability of deploying sophisticated machine-learning models on microcontrollers powered by coin cells, broadening their applicability in various settings for object detection. Employing int8 quantization, our model attains a notable improvement in latency by 45%, coupled with a 70% reduction in RAM consumption and a 65% decrease in flash storage. This research delves into the significance of optimization in the process of choosing latency-efficient deep neural network (DNN) models for various edge computing configurations, emphasizing the delicate balance between hardware capabilities and optimization approaches. Subsequent research efforts will concentrate on refining quantization algorithms to further mitigate the precision differential in computational models.

Downloads

Download data is not yet available.

References

Choi, K.-H. and Ha, J.-E., 2022. Results Comparison of Applying Object Detection and Object Segmentation Methods on Visual Surveillance Dataset.

Li, Y., Jia, J., Zuo, Y. and Zhu, W., 2023. Effective out-of-distribution Detection for TinyML. Available at: https://doi.org/10.1109/icassp49357.2023.10094746.

Krishna, A., Pendkar, N., Kasar, S., Mahind, U. and Desai, S., 2021. Advanced Video Surveillance System. In: 2021 3rd International Conference on Signal Processing and Communication (ICPSC), IEEE.

Che, R., Wang, L., Wang, Y., Lin, Q. et al., 2019. Research on Intelligent Video Surveillance System in Remote Area Based on NB-IoT. In: Proceedings of the 2019 2nd International Conference on Algorithms, Computing and Artificial Intelligence. ACM.

Solunke, B.R. and Gengaje, S.R., 2023. Traditional and Deep Learning based Object Detection Methods. In: International Conference on Emerging Smart Computing and Informatics (ESCI). Available at: https://doi.org/10.1109/esci56872.2023.10099639.

Benali, A. and Amrouch, M., 2023. Object Detection using Deep Learning, CNNs and Vision Transformers: A Review. IEEE Access, 11, pp.35479-35516.

Shivappriya, S., Jasmine, M., Priyadarsini, P. et al., 2021. Cascade Object Detection and Remote Sensing Object Detection Method Based on Trainable Activation Function. Remote Sensing.

Shahid, K., Zhang, Y., Yin, S. et al., 2020. A brief review and challenges of object detection in optical remote sensing imagery. Multiagent and Grid Systems.

Rusci, M., Fariselli, M., Capotondi, A. et al., 2020. Leveraging Automated Mixed-Low-Precision Quantization for Tiny Edge Microcontrollers. Communications in Computer and Information Science, 1325, pp.296-308.

Dokic, K., Martinovic, M., Mandusic, D. et al., 2020. Inference speed and quantisation of neural networks with TensorFlow Lite for Microcontrollers framework. In: 2020 5th South-East Europe Design Automation, Computer Engineering, Computer Networks and Social Media Conference (SEEDA-CECNSM), IEEE, pp.1-6.

Velasco-Montero, D., Jorge, et al., 2018. Performance analysis of real-time DNN inference on Raspberry Pi. Real-Time Image and Video Processing, 2018.

Lokhande, H.N. and Ganorkar, S.R., 2020. Challenges in Scene Interpretation for Video Surveillance. Test-Engineering and Management, 83(2).

Avellaneda, D., Mendez, D. and Fortino, G., 2023. A TinyML Deep Learning Approach for Indoor Tracking of Assets. Sensors, 23(3), 1542. Available at: https://doi.org/10.3390/s23031542.

Solunke, B., Sachin, R. and Gengaje, 2023. A Review on Traditional and Deep Learning based Object Detection Methods. International Conference on Emerging Smart Computing and Informatics (ESCI).

Gayathri, M.D., Lakshmanan, V. and Krishna, 2023. Object Surveillance Detection. International Journal For Science Technology And Engineering. Available at: https://doi.org/10.22214/ijraset.2023.49595.

Khedkar, M., 2021. Wireless Intruder Detection System for Remote Locations. Turkish Journal of Computer and Mathematics Education (TURCOMAT), 12(12), pp.1390-1401. Available at: https://doi.org/10.17762/turcomat.v12i12.7621.

Zhang, L., Wang, Y. and Huo, Y., 2021. Object Detection in High-Resolution Remote Sensing Images Based on a Hard-Example-Mining Network. IEEE Transactions on Geoscience and Remote Sensing.

Gautam, A., Sukhwinder, and Singh, 2021. Deep Learning Based Object Detection Combined with Internet of Things for Remote Surveillance. Wireless Personal Communications.

Chugh, U., Mitra, A., Ankur, et al., 2021. In: IEEE International Symposium on Circuits and Systems (ISCAS), IEEE, pp.1-5.

Sonali, M., Diware, A., Shweta, and Iskande, 2012. Remote Surveillance System for Mobile Application. Network and Complex Systems, 3, pp.14-19.

Amira, S., Mahmoud, M., Sayed, et al., 2020. Object Detection Using Adaptive Mask RCNN in Optical Remote Sensing Images. International Journal of Intelligent Engineering and Systems.

Yanmao, M., Li, M., Ryan, M. and Gerdes, 2023. Remote Perception Attacks against Camera-based Object Recognition Systems and Countermeasures. ACM Transactions on Cyber-Physical Systems.

EL-Hadad, R., Tan, Y.-F, Tan, W.-N, 2022. Anomaly Prediction in Electricity Consumption Using a Combination of Machine Learning Techniques. International Journal of Technology. Volume 13(6), pp. 1317-1325

Hadidi, R., Cao, J., Michael, S. et al., 2019. Robustly Executing DNNs in IoT Systems Using Coded Distributed Computing. In: Proceedings of the 56th Annual Design Automation Conference, ACM, pp.1-2. Available at: https://doi.org/10.1145/3316781.3322474.

Rusci, M., Capotondi, A., Conti, F. et al., 2018. Work-in-Progress: Quantized NNs as the Definitive Solution for Inference on Low-Power ARM MCUs?. In: International Conference on Hardware/Software Codesign and System Synthesis (CODES+ISSS), pp.1-2. Available at: doi:10.1109/CODESISSS.2018.8525915.

Chen, X., Guo, A., Si, X., Yang, J. et al., 2022. A Quantization Model Based on a Floating-point Computing-in- Memory Architecture. In: IEEE Asia Pacific Conference on Circuits and Systems (APCCAS), pp.493-496.

Shirpour, M., Khairdoost, N., Michael, et al., 2023. Traffic Object Detection and Recognition Based on the Attentional Visual Field of Drivers. IEEE Transactions on Intelligent Vehicles.

Kantham, L., Durgam, R., Jatoth, K. et al., 2023. Real-time Classification of Vehicle Logos on Arduino Nano BLE using Edge Impulse. In: 4th International Conference on Signal Processing and Communication (ICSPC), pp.316-320.

Lokhande H.N., Ganorkar S. R., 2021. Masked Face Detection And Voice Based Alert System For Blind People Under Covid-19. International Journal of Engineering Development and Research, 9, pp.104-108.

Mellit, A., Nicola, A., Blasuttigh, P., Massi, P. et al., 2023. TinyML for fault diagnosis of Photovoltaic Modules using Edge Impulse Platform. In: 2023 11th International Conference on Smart Grid (icSmartGrid), pp.1-5.

Son, N. and Truong, S.N., 2020. A Low-cost Artificial Neural Network Model for Raspberry Pi. Engineering Technology & Applied Science Research, 10(2), pp.5466-5469.

Ravi, K. and Jatoth, 2023. Real-time classification of haze and non-haze images on Arduino Nano BLE using Edge Impulse. In: 4th International Conference on Computing and Communication Systems (I3CS), pp.1-7.

Ali, W., Daher, A., Rizik, M. et al., 2021. Porting Rulex Software to the Raspberry Pi for Machine Learning Applications on the Edge. Sensors.

Aftab, F., Bazai, S.U., Marjan, S., Baloch, L., Aslam, S., Amphawan, A., Neo, T.-K., 2023. A Comprehensive Survey on Sentiment Analysis Techniques. International Journal of Technology. Volume 14(6), pp. 1288-1298

Ng.C.H., Connie, T., Choo, K.Y., Goh, M.K.O., 2022. Fusion of Visual and Audio Signals for Wildlife Surveillance. International Journal of Technology. Volume 13(6), pp. 1213-1221

Downloads

Published

26.03.2024

How to Cite

Harshad Lokhande. (2024). Optimizing Real-Time Object Detection on Edge Devices: A Transfer Learning Approach. International Journal of Intelligent Systems and Applications in Engineering, 12(21s), 3896 –. Retrieved from https://ijisae.org/index.php/IJISAE/article/view/6161

Issue

Section

Research Article