Related Experiment Video
Updated: Jul 7, 2025
08:25
Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment
Published on: May 7, 2019
9.0K
3D Object Detection From Images for Autonomous Driving: A Survey.
IEEE Transactions on Pattern Analysis and Machine Intelligence
|December 25, 2023
Summary
This paper surveys image-based 3D object detection for autonomous driving, organizing over 200 recent studies. It analyzes common methods, proposes new taxonomies, and discusses future research directions.
Area of Science:
- Computer Vision
- Robotics
- Artificial Intelligence
Background:
- 3D object detection from images is crucial for autonomous driving.
- Deep learning has driven significant progress in this field since 2015.
- A comprehensive survey of recent advancements is currently lacking.
Purpose of the Study:
- To provide the first comprehensive survey of image-based 3D object detection.
- To organize and analyze the vast body of research from 2015-2021.
- To propose new taxonomies for systematic review and fair comparison of methods.
Main Methods:
- Comprehensive literature review of over 200 works (2015-2021).
- Analysis of common pipelines and their components in image-based 3D detection.
- Development of two novel taxonomies to categorize existing methods.
Main Results:
- Summarization of key theories, algorithms, and applications in the field.
- Systematic organization of state-of-the-art methods using proposed taxonomies.
- Identification of current challenges and future research avenues.
Conclusions:
- The survey fills a critical gap by consolidating knowledge on image-based 3D object detection.
- The proposed taxonomies facilitate a structured understanding and comparison of methods.
- The analysis of challenges and future directions guides ongoing research in autonomous driving technology.
