Now that we’ve shown how HMMs work, let’s provide some more tips on how to improve them. >> Okay, in our example of using HMMs to distinguish between the signs I versus we, we used delta y as a feature. But in reality delta x would be a better feature. >> That’s true, but for other signs delta y would be a good feature. Another good feature is the size of the hand which helps us get some information as to if the hand is coming towards the camera or away. >> Another good feature might be the angle the hand makes to the horizontal. >> So sign language is two handed, we should have these features both to the right and left hands. Now that we have eight features, we are tracking for time frame, how do we integrate it into our models? >> Actually, it’s pretty easy. We just add more dimensions for the output probabilities. All our training and recognition work like before. We just have to calculate multi-dimensional distances instead of using just one dimension. >> You’re right. It gets hard to show on our graphs here but it’s easy enough to code.