Conference Paper2025

Resolution, Compression, and GAN-Enhanced Evaluation of YOLO Models on the Poribohon-BD Vehicle Dataset

Sabrina Masum Meem, Mohammed Tahmid Hossain, Sami Rashid, Ashraful Islam, Rashedur Rahman, Istiaq Ahmed, Asif Mahmud, Saadia Binte Alam, M Ashraful Amin

2025 IEEE 23rd Student Conference on Research and Development (SCOReD)

IEEE, pp. 1-6

  • Sustainable Cities and Communities — Relevance: 67%
DOI: 10.1109/scored68498.2025.11399224URL

Abstract

Object detection in resource-constrained environments poses challenges for real-time intelligent transportation systems (ITS), where low resolution, compression artifacts, and bandwidth limitations hinder performance. This study evaluates YOLOv8, YOLOv10, and YOLOv12 on the Poribohon-BD dataset of 15 native Bangladeshi vehicle classes using three dataset variants: resized images, JPEG-compressed images (Q50, Q75), and GAN-enhanced images restored with Real-ESRGAN. The results show that resolution is the dominant factor, and YOLOv8 delivers the most stable performance under all conditions. YOLOv10 generalized slightly better at very low resolutions, while YOLOv12 was competitive at higher resolutions under compression and GAN restoration but degraded at lower scales. GAN enhancement improved accuracy only when there was sufficient base resolution, revealing a gap between perceptual fidelity and detection accuracy. Overall, these findings provide a systematic framework for selecting YOLO architectures and preprocessing pipelines under degraded inputs, supporting deployment in real-world ITS.

Keywords

  • YOLOv12
  • YOLOv10
  • Real-ESRGAN
  • intelligent transportation systems
  • YOLOv8
  • resource-constrained environments
  • vehicle detection
  • Image resolution
  • JPEG compression

CCDS Authors

Ashraful Islam, PhD
Ashraful Islam

Ashraful Islam, PhD

  • Center Director, CCDS
  • Director, HCI Wing

Published at

CHI, UIST, UbiComp, IJCNN, ICIP, HCII, ICDMW, BLP, Pattern Recognition, IEEE Internet of Things Journal, JMIR, JMIR Human Factors, JMIR Formative Research, IEEE Access, Smart Agricultural Technology, Diabetes Research and Clinical Practice, PLOS ONE

Md Rashedur Rahman, D. Eng.
Md. Rashedur Rahman, D. Eng.

Md Rashedur Rahman, D. Eng.

  • Co-Director, MIRA & NEST Wings

  • Assistant Professor

    Department of Computer Science and Engineering

    Independent University, Bangladesh (IUB)

Published at

CHI, HCII, Scientific Reports, Cognitive Computation, Neural Computing and Applications

Istiaq Ahmed
Istiaq Ahmed

Istiaq Ahmed

  • Research Assistant, HCI Wing

Published at

CHI, HCII

Asif Mahmud
Asif Mahmud

Asif Mahmud

  • Co-Director, IPT Wing

  • Lecturer

    Department of Computer Science and Engineering

    Independent University, Bangladesh (IUB)

Saadia Binte Alam, PhD
Saadia Binte Alam, PhD

Saadia Binte Alam, PhD

  • Director, MIRA & WiSE Wings

  • Associate Professor

    Department of Computer Science and Engineering

    Independent University, Bangladesh (IUB)

Published at

HCII, Scientific Reports, Cognitive Computation, Neural Computing and Applications

Prof. M Ashraful Amin, PhD
M Ashraful Amin, PhD

Prof. M Ashraful Amin, PhD

  • Founder Director, CCDS
  • Director, AI/ML, HCI & DS Wings

Published at

ACL, CHI, UbiComp, ECAI, WACV, PAKDD, IJCNN, ICPR, ICIP, HCII, CLEF, BLP, Scientific Reports, Pattern Recognition, JMIR, IEEE Access, Smart Agricultural Technology, Frontiers in Computational Neuroscience, Diabetes Research and Clinical Practice, PLOS ONE

References

  1. 1.Juan Terven, Diana-Margarita Córdova-Esparza, Julio-Alejandro Romero-González. (2023). A Comprehensive Review of YOLO Architectures in Computer Vision: From YOLOv1 to YOLOv8 and YOLO-NAS. Machine Learning and Knowledge Extraction, 5(4), 1680–1716[10.3390/make5040083]
  2. 2.Ao Wang, Hui Chen, Lihao Liu, Kai Chen, Zijia Lin, Jungong Han, Guiguang Ding. (2024). YOLOv10: Real-Time End-to-End Object Detection. , 107984–108011[10.52202/079017-3429]
  3. 3.Ayoub Benali Amjoud, Mustapha Amrouch. (2023). Object Detection Using Deep Learning, CNNs and Vision Transformers: A Review. IEEE Access, 11, 35479–35516[10.1109/access.2023.3266093]
  4. 4.Harriet L. Dawson, Olivier Dubrule, Cédric M. John. (2022). Impact of dataset size and convolutional neural network architecture on transfer learning for carbonate rock classification. Computers & Geosciences, 171, 105284[10.1016/j.cageo.2022.105284]
  5. 5.Muhammad Azhad Bin Zuraimi, Fadhlan Hafizhelmi Kamaru Zaman. (2021). Vehicle Detection and Tracking using YOLO and DeepSORT. , 23–29[10.1109/iscaie51753.2021.9431784]
  6. 6.Yibo Sun, Zhe Sun, Weitong Chen. (2024). The evolution of object detection methods. Engineering Applications of Artificial Intelligence, 133, 108458[10.1016/j.engappai.2024.108458]
  7. 7.Refaat Mohammad Alamgir, Ali Abir Shuvro, Mueeze Al Mushabbir, Mohammed Ashfaq Raiyan, Nusrat Jahan Rani, Md. Mushfiqur Rahman, Md. Hasanul Kabir, Sabbir Ahmed. (2022). Performance Analysis of YOLO-based Architectures for Vehicle Detection from Traffic Images in Bangladesh. , 982–987[10.1109/iccit57492.2022.10055758]
  8. 8.Shaif Choudhury, Soummyo Priyo Chattopadhyay, Tapan Kumar Hazra. (2017). Vehicle detection and counting using haar feature-based classifier. , 106–109[10.1109/iemecon.2017.8079571]
  9. 9.Md Mahibul Hasan, Zhijie Wang, Muhammad Ather Iqbal Hussain, Kaniz Fatima. (2021). Bangladeshi Native Vehicle Classification Based on Transfer Learning with Deep Convolutional Neural Network. Sensors, 21(22), 7545[10.3390/s21227545]
  10. 10.Farid, Proshanta Kumer Das, Monirul Islam, Ebna Sina. (2025). Bangladeshi Vehicle Classification and Detection Using Deep Convolutional Neural Networks With Transfer Learning. IEEE Access, 13, 26429–26455[10.1109/access.2025.3539713]
  11. 11.Shuo Han, Bo Mo, Jie Zhao, Junwei Xu, Shizun Sun, Bo Jin. (2024). High-Quality Image Compression Algorithm Design Based on Unsupervised Learning. Sensors, 24(20), 6503[10.3390/s24206503]
  12. 12.Soumyajit Gayen, Sourajit Maity, Pawan Kumar Singh, Ram Sarkar. (2024). SimSANet: a simple sequential attention-aided deep neural network for vehicle make and model recognition. Neural Computing and Applications, 37(1), 319–339[10.1007/s00521-024-10480-z]
  13. 13.Naima Islam, Sajeeb Kumar Ray, Md. Anwar Hossain, Mirza AFM Rashidul Hasan, Alamin, Md. Basim Al Zabir Shammo. (2024). Vehicle Classification and Detection Using YOLOv8: A Study on Highway Traffic Analysis. , 1–4[10.1109/icrpset64863.2024.10955913]
  14. 14.Siraj Us Salekin, Md. Hasib Ullah, Abdullah Al Ahad Khan, Md. Shah Jalal, Huu-Hoa Nguyen, Dewan Md. Farid. (2023). Bangladeshi Native Vehicle Classification Employing YOLOv8. Communications in computer and information science, 185–199[10.1007/978-981-99-7649-2_14]
  15. 15.Tanvir Ahmed, Farzana Sharmin Mou, Abu Anas Ibn. Samad. (2023). A Deep Learning based Bangladeshi Vehicle Classification using Fine-Tuned Multi-class Vehicle Image Network (MVINet) Model. , 1–6[10.1109/ncim59001.2023.10212619]
  16. 16.Md Fahad Monir, Md Junayed Hossain, Mohammad Barkatullah, Md Mozammal Hoque, Zoheb Hassan, Tarem Ahmed. (2024). Enhanced Vehicle Detection by Optimized Image Compression in NextG Wireless Network Autonomous Vehicles System. , 1–5[10.1109/vtc2024-fall63153.2024.10757788]
  17. 17.Muhammad Liakat Ali, Topu Biswas, Shahin Akter, Mohammed Farhan Jawad, Hadaate Ullah. (2024). Real‐time object detection for unmanned vehicles in Bangladesh: Dataset, implementation and evaluation. The Journal of Engineering, 2024(12)[10.1049/tje2.70033]
  18. 18.Shaira Tabassum, Sabbir Ullah, Nakib Hossain Al-nur, Swakkhar Shatabda. (2020). Poribohon-BD: Bangladeshi local vehicle image dataset with annotation for classification. Data in Brief, 33, 106465[10.1016/j.dib.2020.106465]
← All publications