Building Computer Vision Projects with OpenCV 4 and C++

The YOLO dataset, vocabulary, and model

Before we start to import the model into our OpenCV code, we have to obtain it through the YOLO website: https://pjreddie.com/darknet/yolo/. This provides pre-trained model files based on the COCO dataset, which contains 80 object categories, such as person, umbrella, bike, motorcycle, car, apple, banana, computer, and chair.

To get all the names of categories and uses for visualization, check out https://github.com/pjreddie/darknet/blob/master/data/coco.names?raw=true.

The names are in the same order as the results of deep learning model confidences. If you want to see some images of the COCO dataset by category, you can explore the dataset at http://cocodataset.org/#explore, and download some of them ...