9th International Conference on ICT for Sustainable Development

Authors - Madhura M. Bhosale, Yogesh S. Angal
Abstract - Object detection is very prominent task in autonomous vehicle domain. Evolution in the area of deep-learning provides great edge for object detection. In this paper we have worked on traffic object detection using YOLO-V8 with model x version on KITTI dataset with adam optimizer, batch size is 8, epoch is 150 and learning rate 0.001. we have used 80% images for training which around 5000 and for remaining 20 % which is around 1500 used for validation. In this experimentation we have used NVIDIA RTX A5000 GPU for model training and we have got mAP accuracy 80%.