Julian Tanke

Hi @una-dinosauria , I just pointed out the visualization as I imagine you chose the joint ids by visual inspection. I'd guess that suggests that the flip is either caused...

Hi, I just take the exp map data taken from [here](http://www.cs.stanford.edu/people/ashesh/h3.6m.zip), as explained [here](https://github.com/una-dinosauria/human-motion-prediction#get-this-code-and-the-data) and then parse the **n x 99** dim into 3D joint coordinates using forward kinematics as...

*Update* The 3D representation extracted from the exponential maps via forward kinematics seems to have some other interesting properties. I align the ground-truth pose (poses from the original dataset and...

@abc123yuanrui @PeikeLi Do you think we can achieve this in an easier way by simple filtering; by replacing https://github.com/PeikeLi/Self-Correction-Human-Parsing/blob/ee5877816acd8cd23be68d8da5661d0931b17427/datasets/simple_extractor_dataset.py#L30 with: ```python allowed_image_extensions = [".png", ".jpg", ".jpeg", ".webm"] self.file_list = list(...

The weight files are lost but re-training the network should not take too much time. You will have to install the [data library](https://github.com/jutanke/pak), download the respective datasets and then simply...

https://github.com/jutanke/person_reid/issues/11

@sowd did you build the image? I think the documentation is a bit lacking in that regard. ``` cd docker && ./build.sh ``` Also, did you add your user to...

We should also smooth the trajectories if the person disappears and reappears on a later frame

You can simply follow https://github.com/jutanke/person_reid/blob/master/sample.ipynb

Just create a settings.txt file with the following setup: ```json { "data_root": "/path/to/your/data" } ```