Transcription: What the key points do are identify how the images correspond with one another. For example, with a car you'd have key points on the wheels and the fenders and the windows and likewise. So, for example, if you put key points on an image but neglected, say, the rear bumper of the cars, Morph would just, in a very uneducated fashion, just blend the two images just like a normal So generally the best way to help with that kind of thing is defining more key points, specifically identifying sections of the image. That way Morph doesn't have to guess and the results in most cases are a lot cleaner. ...
(.mp4)