Optimization of MobileNet SSD Using Pruning, Quantization, and Transfer Learning for Real-Time Vehicle Detection in IoT-Based Security Systems

Afit  Miranto; Purwono Prasetyawan; Iqbal  May Aryanto

doi:10.52436/1.jutif.2026.7.2.5488

Authors

Afit Miranto Department of Electrical Engineering, Institut Teknologi Sumatera, Lampung, Indonesia
Purwono Prasetyawan Department of Electrical Engineering, Institut Teknologi Sumatera, Lampung, Indonesia
Iqbal May Aryanto Department of Electrical Engineering, Institut Teknologi Sumatera, Lampung, Indonesia

DOI:

https://doi.org/10.52436/1.jutif.2026.7.2.5488

Keywords:

Internet Of Things (IoT), MobileNet SSD, Object Detection, Optimization, Real-Time Detection

Abstract

Security is a critical requirement in modern public and private environments, especially in systems that rely on resource-constrained IoT devices. This research aims to optimize the MobileNet SSD (Single Shot MultiBox Detector) model to achieve fast and reliable real-time vehicle and human detection on low-power hardware. The proposed optimization pipeline integrates three techniques: pruning to reduce network redundancy, quantization to accelerate inference and decrease memory usage, and transfer learning using six relevant object classes (person, car, motorcycle, bicycle, bus, and truck). Experiments were conducted on a Raspberry Pi 5 equipped with a camera and local dashboard interface. The optimized MobileNet SSD v2 model achieved a mean Average Precision (mAP) of 0.724 and mAP@0.5 of 0.951, while improving inference speed from 21 FPS to over 24 FPS. These results indicate a balanced trade-off between accuracy, speed, and resource efficiency, enabling stable real-time performance on constrained IoT platforms. The findings contribute to the body of knowledge in embedded and edge AI by demonstrating how integrated model-level optimization can significantly enhance deep learning inference on low-power systems, offering scientific and practical implications for smart surveillance and intelligent traffic monitoring.

Downloads

Download data is not yet available.

References

M. Satyanarayanan, “The Emergence of Edge Computing,” Computer (Long Beach Calif), vol. 50, no. 1, pp. 30–39, 2017, doi: 10.1109/MC.2017.9.

W. Shi, J. Cao, Q. Zhang, Y. Li, and L. Xu, “Edge Computing: Vision and Challenges,” IEEE Internet Things J, vol. 3, no. 5, pp. 637–646, 2016, doi: 10.1109/JIOT.2016.2579198.

F. Bonomi, R. Milito, J. Zhu, and S. Addepalli, “Fog Computing and Its Role in the Internet of Things,” in Proceedings of the First MCC Workshop on Mobile Cloud Computing, 2012, pp. 13–16. doi: 10.1145/2342509.2342513.

L. Qi, X. Zhang, W. Dou, and Q. Ni, “A Survey on Edge Computing in the Internet of Things,” IEEE Access, vol. 8, pp. 195555–195572, 2020, doi: 10.1109/ACCESS.2020.3032861.

W. Liu et al., “SSD: Single shot multibox detector,” Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9905 LNCS, pp. 21–37, 2016, doi: 10.1007/978-3-319-46448-0_2.

J. Redmon and A. Farhadi, “YOLOv3: An incremental improvement,” arXiv preprint arXiv:1804.02767, 2018.

M. Tan, R. Pang, and Q. V Le, “EfficientDet: Scalable and efficient object detection,” Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 10781–10790, 2020, doi: 10.1109/CVPR42600.2020.01079.

M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L.-C. Chen, “MobileNetV2: Inverted Residuals and Linear Bottlenecks,” in CVPR, 2018, pp. 4510–4520. doi: 10.1109/CVPR.2018.00474.

A. Howard and others, “Searching for MobileNetV3,” in ICCV, 2019, pp. 1314–1324. doi: 10.1109/ICCV.2019.00140.

X. Zhang, X. Zhou, M. Lin, and J. Sun, “ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices,” in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2018, pp. 6848–6856. doi: 10.1109/CVPR.2018.00716.

R. Zhang, L. Sun, and Y. Huang, “Improving Embedded Object Detection using MobileNet SSD,” Sensors, vol. 22, no. 15, p. 5654, 2022, doi: 10.3390/s22155654.

S. Han, H. Mao, and W. J. Dally, “Deep compression: Compressing deep neural networks with pruning, trained quantization and Huffman coding,” arXiv preprint arXiv:1510.00149, 2015.

H. Li, A. Kadav, I. Durdanovic, H. Samet, and H. P. Graf, “Pruning Filters for Efficient Convolutional Neural Networks,” arXiv preprint arXiv:1608.08710, 2016, [Online]. Available: https://arxiv.org/abs/1608.08710

B. Jacob, S. Kligys, B. Chen, and et al., “Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference,” in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2018, pp. 2704–2713. doi: 10.1109/CVPR.2018.00286.

Y. Wu, C. Huang, and X. Liu, “Quantization for Deep Neural Networks: A Survey,” IEEE Access, vol. 8, pp. 217123–217136, 2020, doi: 10.1109/ACCESS.2020.3040789.

T. Liang, J. Lu, and X. Chen, “Pruning and Quantization for Deep Neural Network Compression: A Survey,” arXiv preprint arXiv:2101.09671, 2021, [Online]. Available: https://arxiv.org/abs/2101.09671

Y. Choi, J. Lee, and S. Park, “Recent Advances in Model Compression and Acceleration,” IEEE Access, vol. 9, pp. 157526–157538, 2021, doi: 10.1109/ACCESS.2021.3131868.

K. Simonyan, A. Vedaldi, and A. Zisserman, “Edge AI: Concepts, Key Technologies and Future Directions,” IEEE Trans Pattern Anal Mach Intell, 2020, doi: 10.1109/TPAMI.2020.3034684.

N. Sreenu and M. J. Durai, “Intelligent Video Surveillance: A Review,” J Big Data, vol. 6, no. 1, pp. 1–27, 2019, doi: 10.1186/s40537-019-0262-0.

T. Sibanda, M. A. Adedoyin, and P. S. Pillay, “Real-Time IoT-Based Vehicle Detection for Smart Transportation Systems,” IEEE Access, vol. 9, pp. 89687–89699, 2021, doi: 10.1109/ACCESS.2021.3087937.

Y. C. Chiu, C. Y. Tsai, M. Da Ruan, G. Y. Shen, and T. T. Lee, “Mobilenet-SSDv2: An Improved Object Detection Model for Embedded Systems,” 2020 International Conference on System Science and Engineering, ICSSE 2020, pp. 0–4, 2020, doi: 10.1109/ICSSE50014.2020.9219319.

S. K. Ghosh and et al., “Energy-Efficient Approximate Edge Inference Systems,” ACM Trans Embed Comput Syst, vol. 22, no. 6, pp. 1–23, 2023, doi: 10.1145/3589766.

U. Iqbal and et al., “Edge-Computing Video Analytics Solution for Automated Waste Monitoring,” Sensors, vol. 22, no. 20, p. 7821, 2022, doi: 10.3390/s22207821.

W. K. H. Lua and et al., “Lightweight CNN-Based Deep Neural Networks Application in Edge Computing and IoT,” J Ambient Intell Humaniz Comput, vol. 13, pp. 2025–2041, 2022, doi: 10.1007/s12652-021-03567-3.

A. G. Howard et al., “Mobilenets: Efficient Convolutional Neural Networks for Mobile Vision Applications,” arXiv preprint arXiv:1704.04861, 2017.

M. A. Haque and others, “Vehicle and Person Detection in Surveillance Video Using Lightweight CNN,” IEEE Access, vol. 10, pp. 48320–48332, 2022, doi: 10.1109/ACCESS.2022.3178492.

E. Wang and others, “Model Compression and Acceleration for Deep Neural Networks: The State of the Art,” IEEE Signal Process Mag, vol. 37, no. 1, pp. 115–129, 2020, doi: 10.1109/MSP.2019.2951811.

A. Yang and others, “Pruning Neural Networks: A Survey,” IEEE Trans Pattern Anal Mach Intell, 2023, doi: 10.1109/TPAMI.2023.3237834.

H. Park and others, “Survey on Model Compression for On-Device Deep Learning,” ACM Comput Surv, vol. 55, no. 8, pp. 1–35, 2023, doi: 10.1145/3527150.

J. Anwar, K. Muhammad, and A. K. Sangaiah, “IoT-Based Intelligent Security Using Mobile Lightweight Detectors,” IEEE Access, vol. 10, pp. 7443–7455, 2022, doi: 10.1109/ACCESS.2022.3149794.

J. Park and S. Lee, “GPU vs Edge-Device Performance Comparison for Object Detection Models,” Sensors, vol. 21, no. 11, p. 3664, 2021, doi: 10.3390/s21113664.

A. Alzubaidi and others, “Real-Time Object Detection on Raspberry Pi 4 Using MobileNet SSD,” Computers, vol. 10, no. 12, p. 155, 2021, doi: 10.3390/computers10120155.

M. H. Rahman and others, “MobileNet-Based Traffic Monitoring on Raspberry Pi 3/4,” IEEE Access, vol. 8, pp. 181860–181870, 2020, doi: 10.1109/ACCESS.2020.3027561.

Raspberry Pi Foundation, “Raspberry Pi 5 Product Page,” 2023.

S. N. Mousavi and others, “Real-Time Multi-Class Vehicle Detection in IoT Surveillance,” J Real Time Image Process, 2022, doi: 10.1007/s11554-021-01164-4.

J. K. Kim and H. J. Lee, “Smart City Surveillance Using Embedded Deep Learning,” Sensors, vol. 21, no. 6, p. 2058, 2021, doi: 10.3390/s21062058.

S. F. H. Khilji and others, “Deep Learning-Based Surveillance for Smart Transportation,” IEEE Access, vol. 10, pp. 50839–50853, 2022, doi: 10.1109/ACCESS.2022.3162404.

D. R. V Krishna and others, “Small-Object Detection Challenges in Security Video,” IEEE Access, vol. 11, pp. 54483–54499, 2023, doi: 10.1109/ACCESS.2023.3280931.

Z. Dhaief and N. El abbadi, “Road Signs Detection Using SSD MobileNetV2,” Karbala International Journal of Modern Science, vol. 10, 2024, doi: 10.33640/2405-609X.3373.

T.-Y. Lin et al., “Microsoft {COCO}: Common Objects in Context,” in European Conference on Computer Vision (ECCV), 2014, pp. 740–755.

Y.-C. Chiu, C.-Y. Tsai, M.-D. Ruan, G.-Y. Shen, and T.-T. Lee, “Mobilenet-SSDv2: An Improved Object Detection Model for Embedded Systems,” in 2020 International Conference on System Science and Engineering (ICSSE), 2020, pp. 1–5. doi: 10.1109/ICSSE50014.2020.9219319.

D. Biswas, H. Su, C. Wang, A. Stevanovic, and W. Wang, “An automatic traffic density estimation using Single Shot Detection (SSD)and MobileNet-SSD,” Physics and Chemistry of the Earth, vol. 110, no. December, pp. 176–184, 2019, doi: 10.1016/j.pce.2018.12.001.

W. Sun, S. Chen, L. Shi, Y. Li, and Z. Lin, “Vehicle Following in Intelligent Multi-Vehicle Systems Based on SSD-MobileNet,” Proceedings - 2019 Chinese Automation Congress, CAC 2019, pp. 5004–5009, 2019, doi: 10.1109/CAC48633.2019.8996181.

Y. C. Chiu, C. Y. Tsai, M. Da Ruan, G. Y. Shen, and T. T. Lee, “Mobilenet-SSDv2: An Improved Object Detection Model for Embedded Systems,” 2020 International Conference on System Science and Engineering, ICSSE 2020, pp. 0–4, 2020, doi: 10.1109/ICSSE50014.2020.9219319.

J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You only look once: Unified, real-time object detection,” in Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, Jun. 2016, pp. 779–788. doi: 10.1109/CVPR.2016.91.

W. Liu et al., “SSD: Single shot multibox detector,” Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9905 LNCS, pp. 21–37, 2016, doi: 10.1007/978-3-319-46448-0_2.

A. Miranto, “Real Time Object Detection Menggunakan Mobilenet-SSD pada Sistem Keamanan Ruangan dengan Bot Telegram Sebagai Notifikasi User,” JELIKU (Jurnal Elektronik Ilmu Komputer Udayana), vol. 13, no. 1, p. 211, Aug. 2024, doi: 10.24843/JLK.2024.v13.i01.p21.