3DiVi Inc. has announced the release of 3DiVi Face SDK 3.17. The new version features enhanced face identification algorithms and a new ProcessingBlock API for Python.
Reduced False Non-Match Rate
Mobile platforms: in more than 2 times (from 0.13 to 0.06 (FMR=1E-7))
Server video-processing: in more than 2 times (from 0.09 to 0.04 (FMR=1E-7))
Expert systems (>1 mln faces in database): in 6 times (0.12 to 0.02 (FMR=1E-8))
With such improvements, the SDK now allows for flawless face identification even on mobile devices with single-core Cortex A53 processors.
In addition to the algorithm updates, 3DiVi has added a new ProcessingBlock API for Python. This new API will provide developers with more opportunities to add face recognition, face tracking, and emotion estimation to their solutions. The documentation for this new block is currently being updated to ensure that developers can make the most of this new functionality.
The 3DiVi Face SDK is widely used in a range of industries, including security, retail, and hospitality. The software enables real-time face detection, tracking, and recognition, and has been integrated into a variety of applications, including access control systems, time and attendance solutions, and customer experience management platforms.