BioCapture Mobile
SDK for contactless fingerprint capture with a smartphone camera, developed with the AIT and integrated at the Austrian police together with IBM and the Ministry of the Interior. Design, architecture and lead development – from neural finger detection to AFIS matching.
- Core SDK in C++ (CMake, Conan) with SDKs for Android and iOS plus plugins for Cordova, Xamarin and Flutter.
- Finger detection with a neural network (SSD MobileNet v2, TensorFlow Lite).
- Presentation attack detection: recognising fake fingers with EfficientNet models and a dedicated test server.
- Image quality: device-specific camera calibration, fingerprint enhancement and capture quality scoring.
- Standards: NIST-format output with WSQ compression, NIST converter plugin.
- Matching via a service based on OpenAFIS (FastAPI, PostgreSQL, gRPC).
- Police integration: working with IBM and the Austrian Ministry of the Interior (BMI) to integrate BioCapture Mobile into the police's mobile systems.
- Protection: licensing, symbol hiding in the iOS SDK, external security review by SBA Research.
