1. When the app is first launched, it allows users to add objects for new words. The LiDAR sensor attached to an iPad recognizes the surrounding environment through environment recognition, enabling the implementation of detailed AR.
2. After that, objects are recognized using MobileNet V2, and the classified label results are attached to the objects in 3D.
3. The app can also detect flat surfaces and load new 3D mesh models into the screen for learning.
4. When an object is selected on the screen, ray tracing is used to detect where the object is in 3D, and a learning window pops up, allowing the user to start learning.
5. During the test phase, all words are replaced with question marks, and the user selects the object they want to test on and manually inputs the correct answer they previously memorized.
6. If the answer is correct, it is displayed along with an effect.