BA.net Face AI is your local face recognition solution. With exclusive dedication to the space and available for integration work and 24x7 tech support.
There are large providers in the face recognition space:
Amazon has re-purposed their acquisition of Orbeus as part of their AWS platform and these computer vision features are not their core focus. The platform is deeply coupled to the AWS ecosystem. Non-AWS users will have a tough time trying to implement Rekognition in their products.
Google Vision API
A small part of Google’s Cloud Platform, with no facial recognition capabilities. Complex pricing, limited by features combos. Suffers from developer perception that, as with other Google services, the Cloud Vision API could easily be discontinued at any time. This reputation derives from Google deprecating numerous SaaS services throughout the years.
Microsoft Face API (FKA ‘Project Oxford’)
Microsoft has a solid face analysis offering for the cloud. But, the service is highly restricted by usage caps and price instability - look out for their "preview pricing", this is due to increase by 50% once generally available. As part of their bigger Cognitive Services platform it also lacks the focus of a dedicated provider. They do have some cute demos.
IBM Watson Visual Recognition API
One small component of Watson’s vast, growing capabilities. No clear focus on face analysis and highly dependant on IBM’s Bluemix cloud platform. IBM Watson’s strategy is focusing away from visual recognition versus its core text based analyses.
Whilst a large player in the government market, Cognetic is a very complicated and expensive product to set-up, use and maintain.
NEC Face Recognition
A division of NEC based in Japan. Like any ultra-large organization, it’s incredibly difficult to get them on the phone or reply to emails. Their products are good, and for years have been considered a major player. That gap has now been closed. They turn away customers who are spending less than 5MM.
Affectiva is a very well funded startup and somewhat established player. They h
ave raised $34MM over 6 years and are tightly linked to advertising conglomerate WPP. With a focus on emotion analysis, they do not offer facial recognition. Their academic roots means you can expect lots of scientific jargon and complicated products.
They primarily operate in China and are known for their inclusion in Lenovo pro
ducts. Their website is hard to use, with inconsistent information - most of which is out of date.
Great for hacking around with computer vision ideas and conducting research. Up to the developer to figure out how to scale it. With poor recognition rates and very few commercial applications, it’s training wheels.
The Face Recognition deep learning model has an accuracy of 99.38% Using the labelled faces in the wild benchmark. This live demo searches for 5 Hollywood actresses. Penelope Cruz, Angelina Jolie, Scarlet Johanssen, Jessica Alba, Kristen Stewart. Biz API
Upload an image to run Face AI on it.
After repeating this step millions of times for millions of images of thousands of different people, the neural network learns to reliably generate 128 measurements for each person. Any ten different pictures of the same person should give roughly the same measurements.
Once the network has been trained, it can generate measurements for any face, even ones it has never seen before!