search

xuanyuzhou98 / SqueezeSegV2

Implementation of SqueezeSegV2, Improved Model Structure and Unsupervised Domain Adaptation for Road-Object Segmentation from a LiDAR Point Cloud
BSD 2-Clause "Simplified" License
233 stars 69 forks source link

where is the domain adaptation part? #6

Open fangjin-cool opened 5 years ago

fangjin-cool commented 5 years ago

In your paper, you add a Geodesic Correlation Alignment and Progressive Domain Calibration to deal with domain adaptation. But I can not find it in your code. Could you provide that part of code?

AdnanBox commented 5 years ago

@xuanyuzhou98 I'm trying to implement the Progressive Domain Calibration part from your work. I feed in only real data (KITTI) to the trained model (reusing weights from the GCA step). At the end i don't execute the train graph, thus updating only the moving mean and variance. However, my results only get worse. Could you guide me in the right direction?

@jingleFun were you able to implement it?

hcyoo93 commented 3 years ago

Hi, @AdnanBox @jingleFun

Were you guys able to implement PDC part?