How to select and identify objects through the computer camera

I built a simple neural network myself,It can identify the type of fruit.
I want to import this model and use it to call the computer’s camera to recognize something.
Achieve the effect of the following pictures,Make it possible to box out objects and label them.
Where should I get the relevant study materials.

Uploading: QQ截图20210227193959.png…

cv2.VideoCapture