Citation :

Spriha Deshpande , 2025. "Shaping Ethical AI: Bias-Free and Context-Aware Object Detection for Safer Systems" ESP International Journal of Advancements in Computational Technology (ESP-IJACT)  Volume 2, Issue 2: 111-125.

Abstract :

This paper presents an ethical and bias-aware framework for object detection in images using a You only look once (YOLO) based deep learning model, integrating reinforcement learning (RL) and Internet of Things (IoT) data for real-time ethical decision-making in autonomous systems. The framework addresses bias concerns in autonomous systems by assigning ethical scores to different object classes (e.g., pedestrians, vehicles) based on predefined risk and ethical factors. The model uses RL to adjust ethical priorities dynamically based on detected objects and environmental factors, such as GPS data, enabling adaptive decision-making. The system is evaluated using performance metrics like False Negative Rate (FNR) and False Positive Score (FPS), visualizing bias and ethical scores across processed images. This approach demonstrates how combining machine learning, deep learning, IoT, and RL can create more fair and responsible object detection systems in safety-critical applications, such as autonomous driving.

References :

[1] Redmon, J., Divvala, S., Girshick, R., & Girshick, R. (2016). You Only Look Once: Unified, Real-Time Object Detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 779–788. DOI: 10.1109/CVPR.2016.91

[2] Geiger, A., Lenz, P., Stiller, C., & Urtasun, R. (2012). Are we ready for Autonomous Driving? The KITTI Vision Benchmark Suite. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 3354–3361. DOI: 10.1109/CVPR.2012.6248070

[3] He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep Residual Learning for Image Recognition. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 770–778. DOI: 10.1109/CVPR.2016.90

[4] Kong, Y., & Neumann, U. (2017). Adaptive Object Detection for Autonomous Vehicles with Reinforcement Learning. IEEE Transactions on Intelligent Transportation Systems, 18(12), 3257–3267. DOI: 10.1109/TITS.2017.2777112

[5] Sheng, M., Li, L., & Liu, Y. (2019). Multi-Sensor Fusion for Object Detection and Tracking in Autonomous Vehicles. Proceedings of the IEEE Intelligent Vehicles Symposium, 552–557. DOI: 10.1109/IVS.2019.8813819

[6] Xie, L., Wang, X., & Zhang, X. (2020). Ethical Decision Making in Autonomous Systems Using Reinforcement Learning. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 50(9), 3520–3532. DOI: 10.1109/TSMC.2020.2975687

[7] Deshpande, Spriha. (2023). Multi-Sensor Data Simulation and Object Detection: Integrating Cameras, LiDAR, Radar, and Depth Estimation for Enhanced 3D Analysis. Journal of Computer Science and Technology Studies, 5(1), 57-73. https://doi.org/10.32996/jcsts.2023.5.1.8.

[8] Lu, X., Zhang, L., & Chien, S. (2018). Dynamic Object Detection and Ethical Prioritization for Autonomous Driving. Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems, 1234–1239. DOI: 10.1109/IROS.2018.8593849

[9] Zhou, X., & Wang, D. (2019). Improved Object Detection for Autonomous Driving Using LiDAR and Cameras. IEEE Transactions on Vehicular Technology, 68(7), 6702–6713. DOI: 10.1109/TVT.2019.2922734

[10] Hossain, M. A., & Reaz, M. B. I. (2020). IoT and Deep Learning for Ethical Decision Making in Autonomous Vehicles. IEEE Access, 8, 10844–10857. DOI: 10.1109/ACCESS.2020.2964179

[11] Sengupta, S., & Sil, A. (2020). Fairness and Bias-Aware Object Detection for Autonomous Vehicles. IEEE Transactions on Artificial Intelligence, 1(2), 123–135. DOI: 10.1109/TAI.2020.2975154

[12] Mohammad, M. I. (2022). Vehicles Image Dataset. Kaggle. Available at: https://www.kaggle.com/datasets/mmohaiminulislam/vehicles-image-dataset

[13] Deshpande, Spriha. (2025). Ethical Bias with Context-Aware Systems. GitHub. Available at: https://github.com/SprihaDeshpande/Ethical-Bias-with-Context-aware-Systems/

[14] Redmon, J., Divvala, S., Girshick, R., & Girshick, R. (2016). You Only Look Once: Unified, Real-Time Object Detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 779–788. DOI: 10.1109/CVPR.2016.91.

[15] Redmon, J., & Farhadi, A. (2018). YOLOv3: An Incremental Improvement. arXiv preprint arXiv:1804.02767. https://doi.org/10.48550/arXiv.1804.02767.

[16] Mnih, V., Kavukcuoglu, K., Silver, D., et al. (2015). Human-level control through deep reinforcement learning. Nature, 518(7540), 529–533. DOI: 10.1038/nature14236.

[17] This seminal paper introduced Deep Q-Networks (DQN), a breakthrough in reinforcement learning, and is often used as the foundation for many applications, including ethical decision-making models.

[18] Sutton, R. S., & Barto, A. G. (2018). Reinforcement Learning: An Introduction (2nd ed.). MIT Press.

[19] This book is a comprehensive resource on the theory and algorithms behind reinforcement learning, providing foundational knowledge that can help inform the development of ethical decision-making models using RL.

[20] Silver, D., Hubert, T., Schrittwieser, J., et al. (2017). Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm. Nature, 550(7676), 354–359. DOI: 10.1038/nature24270.

Keywords :

Ethical AI, Bias-Aware Object Detection, YOLOv3, Reinforcement Learning (RL), Internet of Things (IoT), Autonomous Vehicles, Ethical Decision-Making, Real-Time Object Detection, Context-Aware Systems, Machine Learning, Deep Learning, Autonomous Systems, Bias Mitigation, GPS Integration, Safety-Critical Systems, Ethical Prioritization.