99 karma · joined May 16, 2017
Less time being bogged down in the quagmire of C++ "features".
However it's not trivial to get it working at the moment, will hopefully soon come back to this project and finish it :D
I rendered all of Estonia from zoom levels 0 to 18 (~16.5M tiles) which is around 20GB and host the tiles on a cheap DigitalOcean instance.
I wrote a realtime human detection library [1] for a robotics project that used HOG + a simple neural net for classification. While it worked okay, I wasn't happy with the precision (around 90%) and decided to try out a simple convnet from Torch (doing the classication on depth images instead of HOG descriptors). The Torch version was slightly slower on a CPU, but both the precision and recall jumped up drastically.
It's a minimal WebRTC server implementation (SDP, STUN, SCTP, DTLS and everything else) contained in a single library. Note that it only implements a subset of the specs to get it working and probably still is buggy (experimental!) as I'm developing it while working on another project that uses it.