Meeting minutes of the weekly meeting 24th May 2017:
==================================
– Martin will get a UCL card when registered for the MD, which will be done soon.
– D. Bouget is happy to run his code with my data and get an acknowledgement.
– Optical flow almost finsished.
– Sync the optical flow code with Loic. This will be done when it is fully tested and working,
we are in touch for this.
– Idea: we could try to show that optical flow + holistic network both working together in real-time
are better and the holistic network alone, particularly with respect to detecting occluded instruments.
A possible way of approaching this problem would be to warp the previous mask and add it to the
merging step of the CNN.
Cheers,
Luis.