- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello everyone.
I am developing a facial recognition access control solution.
The camera model I bought (not yet arrived) is the D435.
My question is what would be the best technique or feature of the camera to identify that the face is real and not a picture of a face?
I read that the facial recognition module (SDK) has been discontinued, is this true?
Thank you all for the attention.
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
In facial recognition, you can differentiate between a real face and a flat photograph by comparing 3D 'landmark' points on a face. The confidence in the accuracy of the identification is increased by comparing more than one image of a particular face.
The 2016 version of the RealSense SDK had a facial recognition system built in, but that SDK is not compatible with the current 400 Series cameras. The facial tech in the 2016 SDK was difficult to program and had bugs though. The link below describes how that old SDK handled face tracking via landmark points.
https://www.intel.co.uk/content/www/uk/en/support/articles/000006099/emerging-technologies.html# facialexpressions Recognition Capabilities of the Intel® RealSense™ Camera
Software called the 'Nuitrack SDK', which is compatible with D435, will soon be adding biometric face recognition. Integrating Nuitrack functionality into your project may be the fastest way to achieve your goal.
https://nuitrack.com/blog# facetracking https://nuitrack.com/blog# facetracking
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page