Your new post is loading...
Strictly speaking, video isn’t an interactive medium, but a new research project from MIT aims to change that: The school’s CSAIL lab has come up with a technique through which viewers can reach out and “touch” objects in videos, manipulating them directly to achieve effects similar to what you’d expect if you were actually touching the object live in the real world.
Basically, that means that using this technique, if you were watching a YouTube video of someone playing guitar and it zoomed in tight on the fretboard, you could theoretically use your mouse to drag across the strings and watch them vibrate as if you’d strummed them in real life. Or, you could even load test an old covered bridge by applying virtual stressors like simulated wind, or a truck rumbling across.
The new CSAIL model works by analyzing vibrations given off by every object, as captured using traditional cameras shooting video that is then analyzed by algorithms developed by the research team. These vibrations, when analyzed by the new technique from as little as five seconds of video of a given object, then provide realistic prediction models that anticipate how the object will react to other movement or forces acting upon them.
Strictly speaking, video isn’t an interactive medium, but a new research project from MIT aims to change that: The school’s CSAIL lab has come up with a technique through which viewers can reach out and “touch” objects in videos, manipulating them directly to achieve effects similar to what you’d expect if you were actually touching the object live in the real world.
Basically, that means that using this technique, if you were watching a YouTube video of someone playing guitar and it zoomed in tight on the fretboard, you could theoretically use your mouse to drag across the strings and watch them vibrate as if you’d strummed them in real life. Or, you could even load test an old covered bridge by applying virtual stressors like simulated wind, or a truck rumbling across.
The new CSAIL model works by analyzing vibrations given off by every object, as captured using traditional cameras shooting video that is then analyzed by algorithms developed by the research team. These vibrations, when analyzed by the new technique from as little as five seconds of video of a given object, then provide realistic prediction models that anticipate how the object will react to other movement or forces acting upon them.