设为首页收藏本站

VALSE

VALSE 首页 活动通知 好文作者面授招 查看内容

20160817-27陈晓智:3D Object Detection for Autonomous Driving

2016-8-12 18:31| 发布者: 程一-计算所| 查看: 7527| 评论: 0

摘要: 报告嘉宾2:陈晓智(清华大学)报告时间:2016年8月17日(星期三)晚21:00(北京时间)报告题目:3D Object Detection for Autonomous Driving 主持人:郑良(悉尼科技大学)报告摘要:In this talk, I will present ...

报告嘉宾2陈晓智(清华大学)

报告时间:2016817日(星期三)晚21:00(北京时间)

报告题目:3D Object Detection for Autonomous Driving

主持人:郑良(悉尼科技大学)


报告摘要

In this talk, I will present a pipeline for 3D object detection in the context of autonomous driving. Our method first aims to generate a set of high-quality 3D object proposals, which are then run through a contextual convolutional network to obtain high-quality object detections. Within this framework, we propose two approaches for 3D object detection. The first approach exploits stereo imagery to place proposals in the form of 3D bounding boxes. We formulate the problem as minimizing an energy function encoding object size priors, ground plane as well as several depth informed features that reason about free space, point cloud densities and distance to the ground. The second approach generate 3D object proposals from a single monocular image. We propose an energy minimization approach that places object candidates in 3D using the fact that objects should be on the ground-plane. We then score each candidate box projected to the image plane via several intuitive potentials encoding semantic segmentation, contextual information, size and location priors and typical object shape. Our experiments show significant performance gains over existing RGB and RGB-D object proposal methods on the challenging KITTI benchmark. Combined with CNN scoring, our approaches achieve state-of-the-art performance on the challenging KITTI benchmark.


参考文献:

[1] Xiaozhi Chen, Kaustav Kunku, Ziyu Zhang, Huimin Ma, Sanja Fidler, Raquel Urtasun. Monocular 3D Object Detection for Autonomous Driving. CVPR, 2016.

[2] Xiaozhi Chen, Kaustav Kunku, Yukun Zhu, Andrew Berneshawi, Huimin Ma, Sanja Fidler, Raquel Urtasun. 3D Object Proposals for Accurate Object Class Detection. NIPS, 2015.


报告人简介:Xiaozhi Chen received the B.S. degree in Electronic Engineering from Tsinghua University, Beijing, China in 2012, where he is currently pursuing the Ph.D degree. He was a visiting student in the Machine Learning Group in University of Toronto in 2015. He is working as a research intern in Autonomous Driving Unit in Baidu Inc. His research interests lie in computer vision and machine learning.


[Slides]

最新评论

Archiver|手机版|小黑屋|Vision And Learning SEminar    

GMT+8, 2020-9-26 22:54 , Processed in 0.030191 second(s), 19 queries .

Powered by Discuz! X3.2

© 2001-2013 Comsenz Inc.

返回顶部