This product uses deep learning based semantic segmentation technology to perform real-time object detection in videos and images. It supports multiple classes of objects and allows users to select the classes of objects to be identified online, such as faces, text, cars, animals, plants, utensils, furniture, etc.