cross-view
cross-view copied to clipboard
[CVPR'21] Projecting Your View Attentively: Monocular Road Scene Layout Estimation via Cross-view Transformation
Hi, I downloaded the data with the provided python from monolayout project. However some of the folders not found. Here are missing folders 1d676737-4110-3f7e-bec0-0c90f74c248f/car_bev_gt 1d676737-4110-3f7e-bec0-0c90f74c248f/road_gt odometry/00/road_dense128 raw/2011_09_26/2011_09_26_drive_0001_sync/ road_dense128 Is there...
Line 174 in file datasets.py : **EDIT** `color_aug = transforms.ColorJitter.get_params(self.brightness, self.contrast, self.saturation, self.hue)` **TO** `color_aug = transforms.ColorJitter(self.brightness, self.contrast, self.saturation, self.hue)`
https://github.com/JonDoe-297/cross-view/blob/957691b0f00bafb59d1fc7146e6408180f3b7649/crossView/CrossViewTransformer.py#L52 https://github.com/JonDoe-297/cross-view/blob/957691b0f00bafb59d1fc7146e6408180f3b7649/crossView/CrossViewTransformer.py#L55 Hi, I was confused about the calculation of cross-view transformer. According to the Figure 4 in your paper, it seems that the "front_x" should be "cross_x" in Line...
According to the paper, there should be a discriminator as well as discriminator loss, but I couldn't find any part of that.
The author provides the following eval command; however, we should use the 1024, 1024, 256 as height, width and occ_map_size if we use the pretrained model provided by the author....
Thanks for your great work. I have the question about the difference between decoder and transform decoder
1. #4 is the first question and definitely it's different from the discription in your paper. I try the two versions and the results seems no different. 2.Magic code: `...
I'm curious about the way to generate panorama HD map. Can you give me some hint?
Thanks for sharing the ideas and codes. Do you also tried to implement position embedding in the cross-view transformer?