Hello, Iām doing a recognition test as a demo project. Once you recognize 5 flat images, you can play each assigned video. However, the 5 images have a similar design with only different names from the 2D characters. So, there is an error where the video of other images is played. I want to improve the design to increase the accuracy of recognition, what method does it take to increase the recognition rate?
Could you post screenshots of your images? Iād also encourage you to check out our documentation which shows examples and characteristics of good and bad image targets.