- Climb recommender. Suggest similar climbs based on input
- Similarity based on beta, holds, ???
- How do we determine if two climbs are similar to each other?
- Body position / beta / keypoint detection thing
- Determine a person's beta based on a video
- Maybe show where and how two betas differ
- Can probably use a pretrained keypoint detector
- Climbing hold segmentation. Given an image of a climb show which holds belong to which routes
- Shouldn't be too bad. Most annoying part is collecting data
- Recommend beta on a route based on an image
- Use keypoint data taken from different images to learn how routes should be climbed?
To edit the project, a virtual environment is recomended. To setup your virtual environment, while in the project directory, run:
$ python -m venv venvTo activate the virtual environment run:
$ source ./venv/Scripts/activateOr in windows
> ./venv/Scripts/activate.batTo install the package in editable mode run:
$ python -m pip install -e .