Learning to Estimate 3D Hand Pose from Single RGB Images

I&#x27;m glad they&#x27;re making inroads here, cause gesture and pose calculation and tracking is awesome.<p>But.. it also seems like more usage of black-box magic math of Neural Networking. I&#x27;m glad it gets results and all, but it just seems.. inelegant. What&#x27;s the algo really doing underneath? Did the algo figure out the joints and their ranges of freedom?<p>The results are spectacular, indeed. But it also seems a bit not like science. It seems more like an oracle system was able to deduce the results - and we&#x27;re nowhere near closer to understanding how to do this. Just, we have a trained system that can.

I think this post would be a lot more informative with the original title &quot;Learning to Estimate 3D Hand Pose from Single RGB Images&quot;<p>The results are astounding for anyone that&#x27;s tried to do similar work in the past. I&#x27;m so pleased that they made the code available.

Interesting.<p>I just finished reading a book called The Hand, by Frank R. Wilson. The author tells how the human hand has shaped our evolution and suggests, among other things, that the development of language came from our hands.<p>A bit heavy at times, but nevertheless warmly recommended.<p><a href="http:&#x2F;&#x2F;www.penguinrandomhouse.com&#x2F;books&#x2F;191866&#x2F;the-hand-by-frank-wilson&#x2F;9780679740476&#x2F;" rel="nofollow">http:&#x2F;&#x2F;www.penguinrandomhouse.com&#x2F;books&#x2F;191866&#x2F;the-hand-by-f...</a>

The vive already has a camera... could this be applied there?<p>What are some other applications?

Okay, but depth cameras exist, full stop, and they obviate the need for an entire GPU worth of processing. It&#x27;s why Uber and Waymo are sinking billions into cheap, solid-state LIDAR. Actual measured sensor data is just so much better.

Here&#x27;s something from six years ago ish with similar results (faster) <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=qok636pe_qw" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=qok636pe_qw</a>

The current title &quot;A 2D Image of your hand and TensorFlow tell you everything&quot; is a little head-scratching. &quot;2D image&quot; as opposed to &quot;stereo&quot; 2D images (or depth-mapped image?). And &quot;tell you everything&quot; makes it sound like some kind of fortune teller.

I don&#x27;t think this is all that amazing.<p>Human fingers only have a very narrow range of motion in a single direction. When the 2D profile of the hand changes, the set of finger movements to arrive at that profile is fairly predictable.