Try live: SSD object detection, Mask R-CNN object detection and instance segmentation, SfMLearner depth and ego motion estimation, directly from your browser!
All these can be run for free with GPU acceleration in Google Colaboratory!
All the code ready to be used in Google Colaboratory is available in this OSSDC.org project:
Click on Google Colaboratory direct links or download as raw the demo Jupyter notebooks listed there, upload them to Google Drive and open them in Colaboratory.
Thanks for the excellent article!
In the SSD demo I also used the best open source SSD object detection, Paul Balança’s TensorFlow implementation.
See here some examples from runs of each model:
The picture above shows a live video streamed from my phone to Google Colaboratory notebook, processed there and streamed back to my laptop in less then 500ms, while doing SSD object detection in 0.06sec/frame, all for free in the cloud with GPU acceleration!
Interested to push forward the Open Source Self Driving Car Initiative, please join us here:
For more updates on OSSDC, Robotics and Artificial Intelligence, follow me on Twitter @gtarobotics and at RaceAI.net
