Scholarlink

Object detection has been a key feature of autonomous driving. Autonomous driving is believed to be the solution to the hike in accidents. To develop an object detection model for an autonomous vehicle in smart cities, a few methods were identified by research and studies. Deep learning algorithm that uses artificial neural networks to replace brain functions can perform sophisticated computations on large amounts of data. From the various methods and algorithms available, the performance of each model will vary for each study. This study aims to investigate and identify the best algorithm for detecting objects in smart cities based on deep learning. The chosen algorithm, You Only Look Once (YOLOv5) is then used to build an object detection model with a driving dataset in a framework. The performance of the model trained will then be evaluated and the results will be analyzed. One of the performance evaluation metrics included in this study is the Mean Average Precision (mAP) which will be compared to a few other object detection models.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Deep Learning Approaches for Object Detection in Autonomous Driving: Smart Cities Perspective

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Othman O. Khalifa,
  • Hariz Naufal Mohd Daud,
  • Elmustafa Sayed Ali,
  • Mamoon M. Saeed

摘要

Object detection has been a key feature of autonomous driving. Autonomous driving is believed to be the solution to the hike in accidents. To develop an object detection model for an autonomous vehicle in smart cities, a few methods were identified by research and studies. Deep learning algorithm that uses artificial neural networks to replace brain functions can perform sophisticated computations on large amounts of data. From the various methods and algorithms available, the performance of each model will vary for each study. This study aims to investigate and identify the best algorithm for detecting objects in smart cities based on deep learning. The chosen algorithm, You Only Look Once (YOLOv5) is then used to build an object detection model with a driving dataset in a framework. The performance of the model trained will then be evaluated and the results will be analyzed. One of the performance evaluation metrics included in this study is the Mean Average Precision (mAP) which will be compared to a few other object detection models.