search

wmcnally / deep-darts

DeepDarts is the first deep learning-based automatic scoring system for steel-tip darts. It predicts dart scores from a single image taken from any camera angle.
115 stars 39 forks source link

How to Detect #4

Open speedoandy opened 1 year ago

speedoandy commented 1 year ago

Could u explain how to detect with any image?

chadha-sridi commented 2 weeks ago

Hi, This is what I did to run it on my images: 1/First add a "configuration file" for your images in deep-darts/configs (one similar to the 2 yaml files already there) 2/Then you need annotate your images by running the annotate.py this will create the labels (you might have to put them in a specific directory: see what you get from the error message). This will create the true "labels" for your images. 3/ make sure to comment the assert line in dataloader.py (asserts the dataset is either d1 or d2) and add a line to handle your data. In predict.py also handle the case for your own config by setting your own value for the labels_path 4/Then you run the predict.py on your images