Our Technology
SDK
OnPrem
Industries
Fintech
Education
Security
Healthcare
Retail
Hospitality
Government
Company
About Us
In the News
Blog
Contact Us
BigBear.ai Trueface SDK
Customize our computer vision solutions into your infrastructure or product.
Written in C++, our developer SDK is optimized for speed and scalability for your performance-critical applications.
Key Features
CPU and GPU support
Optimized for performance
API in C++, Python, Java (Android) and Swift (iOS)
Deployable anywhere
Lightweight and high precision model options for any deployment
Able to operate completely offline
Current Capabilities
Face recognition
(1:1 and 1:N face matching)
Face extraction and
landmark detection
Object detection
Liveness
Head pose estimation
Supported Architectures
x86, ARM (AArch64, AArch32)
Custom (on request)
Supported Operating Systems:
Supported Architectures
x86
ARM
Supported Operating Systems:
CPU Benchmarks
i9-11900H
AWS C6I.2XLARGE (X86)
Ram Req.
Preprocess Image
4.6 ms
6.4 ms
-
Face and Landmark Detection
7.9 ms
12.5 ms
42 MB
Face Recognition Template Extraction TFV7
1 to N Identification Search (N = 1,000,000) TFV7
67.2 ms
26.9 ms
94.2 ms
21.4 ms
500 MB
2250 bytes / template
Object Detection
29.3 ms
39.8 ms
70 MB
Passive Spoof Detection
24.1 ms
53.3 ms
140 MB
GPU Benchmarks
AWS G4DN.XLARGE (T4)
VRAM Req.
Face and Landmark Detection
1.8 ms
130 MB
Face recognition Template Extraction
TFV6, batch size = 1
3.2 ms
332 MB
Face recognition Template Extraction
TFV7, batch size = 1
Object Detection
6.0 ms
6.3 ms
430 MB
160 MB
View all Benchmarks
SDK Resources
SDK documentation
Sample apps
BigBear.ai Trueface SDKs
BigBear.ai Trueface API
