Computer Vision Explorations

Darknet YOLO and Tiny YOLO: https://pjreddie.com/darknet/yolo/
This is the James Bond movie.  This set of libraries can be deployed on a Raspberry Pi

Fast CV SDK: https://developer.qualcomm.com/software/fastcv-sdk
This is an open source free license which is used for mobile devices library set by Qualcomm which we are evaluating in Android.  This has inbuilt libraries for object detection.  If the performance is good we could use this as one of the options in an android environment.  We could get this running in Android Things OS and see the results

Active Vision: https://www.youtube.com/watch?v=InqV34BcheM
This is another set of libraries that is extensively used for surveillance cameras which we are trying out

Owal.io: http://www.owal.io
Artificial Intelligence that transforms every camera into a sentinel, doing the work for you.  They convert squandered video information into transparent, intelligent presence for your building and communities

Open Pose Library: https://github.com/CMU-Perceptual-Computing-Lab/openpose
This is extensive library set of different poses of human body which is used as a base reference to detect and count people.  The computing power required here is high, however we are trying out different way to optimize the performance

Body Estimation