melisandeteng

Would it be possible to add remote sensing from MDPI please ? https://www.mdpi.com/journal/remotesensing Thank you!

Ok @vict0rsch ! I'll update the `resources` page for the datasets, and I created a new repo for all Simulated dataset related utils : https://github.com/cc-ai/SimData GitHubcc-ai/SimDataAbout our simulated world, how...

We used [Deeplabv3+MobileNet](https://github.com/cc-ai/Deeplabv3) model to generate segmentation maps for our streetview dataset of 1950 images available at `/network/tmp1/ccai/data/munit_dataset/non_flooded/streetview_mvp`. This is a reduced version of our real dataset on which depth...

It seems that Deeplabv3 does not do such a great job at segmenting sky on Streetview images, especially when there are clouds (there is a lot of warping in the...

Segmentation maps computed with HRNet model are available at `/network/tmp1/ccai/data/munit_dataset/non_flooded/streetview_mvp_seg_HRNet`for the cityscapes labels and `/network/tmp1/ccai/data/munit_dataset/non_flooded/streetview_mvp_seg_HRNet_simlabels`for the labels (merged from cityscapes) that correspond to our simulated dataset labels.

**Deeplab versions** Currently for the ground segmentation in images that we are using Deeplab v2 pretrained on Cityscapes. The only available Deeplab v3 pretrained models were trained on Pascal VOC...

**Deeplab v2 vs Deeplab v3+MobileNet** [Code in the Deeplabv3 repo]( https://github.com/cc-ai/Deeplabv3.git) We tested a Deeplabv3+MobileNet pretrained on Cityscapes (see [this repo](https://github.com/VainF/DeepLabV3Plus-Pytorch) on 30 images of Canadian landmarks. DeeplabV3+mobilenet seems to...

The last one is sort of a failure case.

We actually need to compare versions of masks generated with deeplab that were computed on images of the same resolution. We compute Deeplabv3 on images resized to 513 px *...

We compare Deeplabv2 and Deeplabv3 on 20 images that we labeled with Labelbox. We infer on the images resized so that the longest side is 600 px. Pixel accuracy: v2:...