How do I avoid having false positives with image tracking? I have 5 images for tracking and each one should be used individually. However, I noticed that sometimes the camera mistakes one image for another even being different in content: each one has a different person.
Iβm also overlaying a video on top of each image, and the other problem i have is that the video doesnβt stop or pauses when the tracking is lost, which ends up conflicting with other videos.
Are you able to share the images you are using, or an example of two that the engine is getting mixed up? The best thing you can do is add distinct feature points to each image. I would only expect to see this behavior for images that look similar.
Have you seen the pause video on lost sample project?