I was trying to create my own working facial detection and recognition system, but now I had a meeting with our director of engineering today, in which out of the blue he told me that they have decided that instead of using OpenCV, they would like to integrate an already existing API for their facial detection system.
So now my job is to explore the various API’s in the market for facial detection and recognition, and test them on the photos in the Truly Madly database.
This sudden change of my project details irks me. But I guess this come as no surprise from a startup.
The image detection and recongtion API’s which I’ve found could be use to us are:-
1. Lambda Labs
2. Skybiometry API
3. Project Oxford, by Microsoft
4. A few API’s on Mashape
I won’t be creating my own image detection code using OpenCV now. My job now is kinda reduced to testing API’s now. I have a database of about 10k images, sorted according to the one’s ‘accepted’ or ‘rejected’ by the Truly Madly promoters for both males and females.
I’ll have to additionally test the rejected images for reasons of rejection, to check how each API i’m testing responds to rejected images in what manner.