search

HKUST-Aerial-Robotics / Stereo-RCNN

Code for 'Stereo R-CNN based 3D Object Detection for Autonomous Driving' (CVPR 2019)
MIT License
686 stars 177 forks source link

Can I adjust this code for normal object detection ? #22

Open Yaoxingtian opened 5 years ago

Yaoxingtian commented 5 years ago

Dear Li: I just wanna detect normal objecst via 3D methods，but not cars,maybe tables or others. could you kindly give some advice? Do you think if I should adjust your code , I feel so confused . Looking for your early reply, thanks so much!

PeiliangLi commented 5 years ago

Hi, the answer is yes. The non-trivial part is that the dense alignment module requires the awareness of the object shape. For the car which has a regular shape, we use the cuboid instead. For general objects, it is feasible to combine with some prior shape or shape recovering module.

Yaoxingtian commented 5 years ago

Hi Li: Thank you ! But how can I make datasets, compare with the dataset including left/right images / calib/ velodyne /labels, if I collect data only includ images and velodyne for the general objects ? And how can I delete calib code in your Stereo-RCNN?

发送自 Outlookhttp://aka.ms/weboutlook

发件人: Peiliang notifications@github.com 发送时间: 2019年5月26日 19:32 收件人: HKUST-Aerial-Robotics/Stereo-RCNN 抄送: Yaoxingtian; Author 主题: Re: [HKUST-Aerial-Robotics/Stereo-RCNN] Can I adjust this code for normal object detection ? (#22)

Hi, the answer is yes. The non-trivial part is that the dense alignment module requires the awareness of the object shape. For the car which has a regular shape, we use the cuboid instead. For general objects, it is feasible to combine with some prior shape or shape recovering module.

― You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/HKUST-Aerial-Robotics/Stereo-RCNN/issues/22?email_source=notifications&email_token=AFCSU4UB4EWYAAY3I575JS3PXLQTZA5CNFSM4HOZUX72YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODWIL6OY#issuecomment-496025403, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AFCSU4RMBX4ETWDGXTRJNETPXLQTZANCNFSM4HOZUX7Q.