3D Object Detection in Adverse Conditions?

utiasASRL / pyboreas

Devkit for the Boreas autonomous driving dataset.

BSD 3-Clause "New" or "Revised" License

90 stars 9 forks source link

3D Object Detection in Adverse Conditions? #22

Closed MartinHahner closed 2 years ago

MartinHahner commented 2 years ago

Hi @keenan-burnett, great dataset! 🎉

Are you planning to release 3D Object Detection labels also for sequences containing rain and snowfall? In your paper you e.g. mention Boreas-Objects-V1, so will there be a planned V2 release?

keenan-burnett commented 2 years ago

Hi Martin!

We're looking into this internally. Since we already have the data, it wouldn't be too much extra work to label some frames from the rain or snow sequences. I can't promise that there will be a V2 with these additions, but I'll say it could happen over the next year.

meiiqi commented 1 year ago

Hi @keenan-burnett, any recent updates on this? Thanks!

keenan-burnett commented 1 year ago

Hi @meiiqi, we have plenty of interesting data taken during adverse weather but I currently don't have enough budget allocated for another Scale data annotation run. I know there are some researchers at my institute who are interested in pursuing this towards their winter autonomous driving research but it won't be happening any time soon.