More info on this website
-
- Clone or download this repository;
-
- Choose and download weights and config files for NN from this website;
-
- Put these files in Temp folder;
-
- pip install -r requirements.txt.
In Base.ipynb set following variables:
- path_to_video: full path to folder with videos;
- temp_folder: full path to Temp folder with (coco.names, labelImg.py, predefined_classes.txt);
- Set names for downloaded weights and config files:
net, outputlayers = load_pretrained_model(temp_folder, 'model.weights', 'model.cfg').
- After starting all cells, in the labelImg program select and save the necessary area and then close program window.
In Jupyter Notebook -> Base.ipynb -> Cell -> Run All