I need a Computer Vision developer / programmer (Intermediate/Advanced level) to:
1. Help me select a SDK (free or paid) for Windows, preferably for .NET, with a set of libraries to perform Face Detection + Face Recognition on live video streams
2. Program the necessary algorithms to accomplish the following tasks:
a) There will be two dedicated cameras connected to the same PC, each providing a video stream (at the lowest needed frame-rate).
b). People will pass in front of camera #1 at time t1. Each face must be detected and a face descriptor stored.
c). The same people will pass in front of camera #2 at time t2, in a different order. All the people that pass in front of camera #2 must also have passed by camera #1 (but might not have been detected). Those who do have a high level of match confidence will get a positive match and be removed from the database; those with a low level of match confidence will be discarded.
VERY IMPORTANT: Most SDKs use a training process for STORAGE and only function in real time for DETECTION. This is not acceptable. The face descriptor creation must ALSO be made in real time, after Face Detection in the first camera video stream.
The source code must be made in C#, C++, VB.NET or Python and commented to be easily changed by others.
OpenCV can be used. Other suggestions may be proposed.
About the recuiterMember since Jul 4, 2017 Grigor G.
from Oberosterreich, Austria