Pembelajaran Kolaboratif Berdasarkan Two-Branch Neural Network dan YOLOv5 Untuk Deteksi Objek Pada Kendaraan Otonom
Abstrak
Abstrak—Seiring dengan kemajuan teknologi dan otomatisasi, perkembangan pada Autonomous Vehicle (AV) meningkat secara signifikan. Object detection memegang peranan penting pada teknologi AV. Pada penerapannya, kondisi cuaca yang buruk mengakibatkan terjadinya penurunan performa sistem dalam mendeteksi objek terutama ketika cuaca berkabut. Tugas Akhir ini menganalisis konfigurasi dari pembelajaran kolaboratif an- tara algoritma dehazing dan object detection untuk meningkatkan kinerja sistem AV dalam mendeteksi objek di kondisi cuaca berkabut. Algoritma dehazing yang digunakan adalah Two- Branch Neural Network, sedangkan algoritma object detection yang digunakan adalah YOLOv5. Pada YOLOv5 dilakukan optimasi dengan hyperparameter tuning untuk mendapatkan nilai pengukuran terbaik. Hasil penelitian menunjukkan bahwa model pembelajaran kolaboratif memiliki mAP yang lebih tinggi dari model YOLOv5 orisinal, dengan nilai 71,5%. Di sisi lain, konfigurasi hyperparameter terbaik didapatkan pada nilai learn- ing rate 0,00334; batch size 32; dan lainnya didapatkan dari hyperparameter VOC. Hal ini meningkatkan mAP dari 71,5% ke 74,8%.
Kata kunci—AV, YOLOv5, two-branch neural network, object detection, image dehazing, hyperparameter
Referensi
F. Munir, S. Azam, M. I. Hussain, A. M. Sheri, and M. Jeon, “Au- tonomous vehicle,” Proceedings of the 2018 International Conference on Sensors, Signal and Image Processing - SSIP 2018, 2018.
S. Zang, M. Ding, D. Smith, P. Tyler, T. Rakotoarivelo, and M. A. Kaafar, “The impact of adverse weather conditions on autonomous vehicles: How rain, snow, fog, and hail affect the performance of a self-driving car,” IEEE Vehicular Technology Magazine, vol. 14, no. 2, p. 103–111, 2019.
B. Li, X. Peng, Z. Wang, J. Xu, and D. Feng, “Aod-net: All-in-one dehazing network,” 2017 IEEE International Conference on Computer Vision (ICCV), 2017.
Y. Yu, H. Liu, M. Fu, J. Chen, X. Wang, and K. Wang, “A two- branch neural network for non-homogeneous dehazing via ensemble learning,” 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2021.
S.-C. Huang, T.-H. Le, and D.-W. Jaw, “Dsnet: Joint semantic learning for object detection in inclement weather conditions,” IEEE Transactions on Pattern Analysis and Machine Intelligence, p. 1–1, 2020.
G. Jocher, A. Stoken, J. Borovec, NanoCode012, A. Chaurasia, TaoXie, L. Changyu, A. V, Laughing, Tkianai, and et al., “ultralytics/yolov5: v5.0 - yolov5-p6 1280 models, aws, supervise.ly and youtube integrations,” Apr 2021. [Online]. Available: https://doi.org/10.5281/zenodo.4679653
C. Sakaridis, D. Dai, and L. V. Gool, “Semantic foggy scene understand- ing with synthetic data,” International Journal of Computer Vision, vol. 126, no. 9, p. 973–992, 2018.
J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, “You only look once: Unified, real-time object detection,” 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016.
M. Kalinina and P. Nikolaev, “Research of yolo architecture models in book detection,” Proceedings of the 8th Scientific Conference on In- formation Technologies for Intelligent Decision Making Support (ITIDS 2020), 2020.
D. Thuan, “Evolution of yolo algorithm and yolov5: The state-of-the-art object detention algorithm,” 2021.
A. Bochkovskiy, C. Wang, and H. M. Liao, “Yolov4: Optimal speed and accuracy of object detection,” CoRR, vol. abs/2004.10934, 2020. [Online]. Available: https://arxiv.org/abs/2004.10934
C.-Y. Wang, H.-Y. M. Liao, Y.-H. Wu, P.-Y. Chen, J.-W. Hsieh, and I.-H. Yeh, “Cspnet: A new backbone that can enhance learning capability of cnn,” 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2020.
F. Hutter, L. Kotthoff, and J. Vanschoren, Automated machine learning: methods, systems, challenges. Springer Nature, 2019.
G. I. Diaz, A. Fokoue-Nkoutche, G. Nannicini, and H. Samulowitz, “An effective algorithm for hyperparameter optimization of neural networks,” IBM Journal of Research and Development, vol. 61, no. 4/5, pp. 9–1, 2017.
D. R. Wilson and T. R. Martinez, “The need for small learning rates on large problems,” in IJCNN’01. International Joint Conference on Neural Networks. Proceedings (Cat. No. 01CH37222), vol. 1. IEEE, 2001, pp. 115–119.
B. D. Hammel, “What learning rate should i use?” Mar 2019. [Online]. Available: http://www.bdhammel.com/learning-rates/
P. M. Radiuk, “Impact of training set batch size on the performance of convolutional neural networks for diverse datasets,” 2017.
S.-H. Gao, M.-M. Cheng, K. Zhao, X.-Y. Zhang, M.-H. Yang, and P. Torr, “Res2net: A new multi-scale backbone architecture,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 43, no. 2, p. 652–662, 2021.
M. Cordts, M. Omran, S. Ramos, T. Rehfeld, M. Enzweiler, R. Be- nenson, U. Franke, S. Roth, and B. Schiele, “The cityscapes dataset for semantic urban scene understanding,” 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016.
