Robo-Refs and the NBA of the Future, Part I

Well, this is an odd confluence of topics! Today’s post is the point at which video games, technology, sports, acting, and future-gazing come together.

Last May I wrote myself an email (what, you don’t do that?) outlining an idea in which the NBA could, using a collection of numerous high-def cameras and some custom software, eliminate “flopping” from the league. Flopping, for those who aren’t familiar with the term, is the act of, well, acting like you’ve been fouled (flailing your arms, falling down, basically exaggerating or, in some cases, inventing “contact”) in order to deceive the referee into making a false call. While certainly not unique to basketball, as numerous soccer-related .gifs can attest, flopping has in recent years become a plague on the sport. The blame for this, of course, lies with the players doing the flopping, not with the officials. The referees are, after all, mere human beings each limited to a single perspective on the action, and as well-trained and hard-working as they are, simply cannot be expected to get every call right all the time.

That’s where the cameras and software come into play. But before we get into that, let’s set things up by taking a relevant detour into the world of home video game consoles.

In 2010, Microsoft released the Kinect, a hardware peripheral for their popular Xbox 360 game console. Intended to facilitate new ways of interacting with games, the Kinect utilizes a microphone, infrared projector, and various sensors to, among other things, track the joints of a human skeleton in real-time. The intended purpose is to allow players to interact with video games using the movement of their bodies, but what makes Kinect noteworthy, are the other things it is being used for in university and commercial laboratories all around the world, from displaying a collection of dresses onto a live video feed of customers at a store in Moscow to allowing researchers in the US to better assess autism in children. And what makes Kinect relevant to this particular discussion about NBA flopping is what the next Kinect (shipping mere months from now) will be able do with its increased resolution and new sensors: in addition to recognizing gestures and detecting a gamer’s heart rate (yes, really,) the new Kinect can track up to six skeletons at once, recognizing not only the rotation of joints, but also the weight put on each limb and the speed of a player’s movements. Not bad for a piece of consumer-grade hardware packed into a box with a game console for $500.

Now, most of you are probably already drawing the connection between motion-sensing software and anti-flopping, but I want to make one more detour along the way before we pull it all together. Next stop: fight scene choreography!

The thing about filming a fight scene, for TV or movies or anything else, is that, in most cases, your actors don’t want to actually get punched in the face and/or thrown across the room. So in the case of a punch, Actor A will swing his or her fist through the air in front of the face of Actor B, and Actor B will adopt a convincing grimace of pain and fling himself or herself backwards as if being propelled by the force of the blow. What you may not have considered, though, is that this fake action really only “works” from one angle, preferably the one the camera is pointing. If you were a spectator on the set of the shoot, watching the same punch from a spot perpendicular to the camera, you’d be able to clearly see the safe distance between fist and face. And, though you might need to see it in slow motion to really break it down, you’d also observe that Actor B’s backward motion was self-inflicted based on the way he/she moved. Envision this in slow-motion: if Actor B truly was punched in the face, the head would be the first to move along the trajectory of Actor A’s punch. Thanks to the bones in the neck and spine, the head would then pull the chest along with it, the movement of the shoulders would cause the hands to swing upward, and so on through the series of connected bones that make up our skeletons. However, in the case of our “staged” punch, since there is no outside force acting on Actor B’s face, all the motion must come from Actor B herself. Even without knowing all of Newton and Euler’s laws of motion, given a good view of the action you can generally spot the difference between a real action and a fake one. That’s why wire-work in movies almost always looks cheesy: the physics are wrong.

Alright, so here we are. We know we’ve got currently-existing, affordable technology that can track human bodies in motion and analyze the forces acting upon them, AND we’ve established that you can’t truly “fake” motion. All that remains is to install a sufficient number of cameras around an NBA arena to ensure good coverage of various angles, and put some motion-analyzing software to work, and you’ve got a flop-proof system. It won’t matter if the referee had a good angle of the alleged infraction, because the program would know almost instantly, based purely on math, whether contact occurred or not, and whether existing contact could have possibly caused the outcome.

Sidenote: It’s worth mentioning that I’m not the only one thinking along these lines. Just a week after I outlined my ideas, Dallas Mavericks owner Mark Cuban funded a university study into determining whether motion-sensing technology can distinguish between actual fouls and flopping. (I figure if I’m independently having the same ideas as a man worth over $2 billion, I must be doing something right.)

The real shock came a few months later when I read that the NBA is already planning to install motion-tracking cameras in every arena for this upcoming season. They’re doing so under the guise of stats-tracking and general data collection, but having these cameras installed solves the hardware problem for our anti-flopping system. All the pieces are in place.

It’s going to be very interesting to see what point the league begins to visibly use the technology to combat flopping. Don’t be surprised if it’s sooner rather than later.

In Part II, coming soon, we’ll look beyond this season at where this technology could take us.


One response to “Robo-Refs and the NBA of the Future, Part I

  1. Pingback: ROBO-REFS AND THE NBA OF THE FUTURE, PART II | thinkface·

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s