Innovatrics OEM Solutions
Through lnnovatrics SDKs our customers are given access to top-ranked fingerprint matching and face recognition algorithms. Thanks to our expertise in porting SDKs, we are able to port our libraries on any hardware. The algorithms make extraction, verification and identification basically instant even on constrained hardware. Their footprint is kept intentionally low to utilize the widest variety of hardware platforms. These features allow our embedded algorithms to work in access control terminals, smartphone fingerprint sensors, small area sensors embedded in credit cards and in many other devices.
The support of ANSI/ISO standards makes devices with embedded SDKs highly interoperable. Even face recognition is now available on compact platforms and in smartphones, making the technology accessible to mainstream applications.
Access Control and Management
Devices with fingerprint, face or iris sensors are ideal targets for Innovatrics OEM algorithms, providing both security and access control functions.
Smart Device Security
Due to their easy portability, Innovatrics algorithms can be easily embedded into smart devices in IoT ecosystems.
Small footprint of Innovatrics algorithms makes them the ideal candidate in cases where storage and performance are scarce.
High accuracy matching even on very small-area fingerprints scanners.
Simple and flexible licensing model.
Tested and adapted for various scanner types and sizes.
10+ years experience in porting our SDKs onto various platforms and OS.
Top-ranked and certified in NIST MINEX III, STQC, NIST FpVTE, NIST PFT II, Intel, and others.
Experience with Porting to these CPUs
• ARM architecture: ARMv3, ARMv4, ARMv4T, ARMv6, ARMv7
• ARM family and core: Cortex M4, Cortex A7, Cortex A53
• Strong ARM
• Samsung S3C2416
• MIPS I & II
• Embedded Linux
• Android 2.2 and higher
• Apple iOS
• Real-time Operating System (RTOS)
• No OS (bare metal)
• Others (analysis required)
Minimum HW Requirements*
• Template Extraction 150 MHz CPU, 400 KB RAM
• Verification 180 MHz CPU, 400 KB RAM
• Identification 600 MHz CPU, 2 MB RAM (per 1000 FP templates)
• Detection, template extraction and verification: 800 MHz CPU, 60 MB RAM
Minimum Flash Memory Requirements*
(Innovatrics SDK library size)
• ANSI&ISO SDK: 650 kB – 1 MB
• IDKit SDK: 2 – 2.5 MB
• IFace SDK: 20 MB
• IDKit Multi: 25 MB
*Even platforms that fail to meet recommended hardware requirements can be ported to. This requires prior consultation by our engineering team.
• ISO template extraction: 565 ms, Verification: 46 ms (ARM Cortex M4 180 Mhz, 256 kB RAM)
• Proprietary template extraction: 65 ms, Identification (1:10000): 105 ms (4xARM Cortex A53 1.2GHz, 1 GB RAM)
• Face detection: 236 ms, Template extraction: 706 ms (Stock Raspberry Pi 3)
• Face detection: 110 ms, Template extraction: 293 ms (Qualcomm Snapdragon: 8x ARM Cortex A53, 1 GB RAM, custom aarch64 Linux OS)