It's Computer Vision project in which what I did I detect a hand and also draw two line first one is safety line and second one is machine line. This safely line alert the user doesn’t cross their hand after this line. But after they ignore this alert alarm and cross that safety line when his hand reached the machine line the machine automically gets stopped. Here I train object detection model taking consideration of both accuracy and time (FPS). Also introduce a lot of variation in dataset by which my model can detect all types of hand, whether in terms of color or gloves and lighting conditions.