Introduction

InspireFace is a powerful, cross-platform face recognition SDK written in C/C++ that enables high-performance facial analysis across a wide range of hardware platforms. Designed for real-world deployment in mobile, embedded, and server-side environments, InspireFace provides a full pipeline for facial processing, from detection to recognition, with support for advanced features such as liveness detection, mask detection, facial attributes, and more.

Core Features

• Face Detection — Fast and accurate face localization in images and video streams.
• Facial Landmarks — High-precision alignment for downstream tasks.
• Face Embeddings & Recognition — Compact feature extraction and identity comparison.
• Face Tracking — Smooth tracking of faces across video frames.
• Mask Detection & Liveness Check — Identify whether a face is masked or spoofed.
• Pose Estimation — Euler angle (roll, pitch, yaw) calculation for each face.
• Face Attribute Analysis — Age, gender, and expression inference.
• Expression & Action Detection — Blink, nod, and head-shake detection for interactive apps.
• Quality Assessment — Image quality metrics to ensure robust inference.

Flexible Deployment

InspireFace supports deployment across a broad set of hardware and platforms:

• CPUs: x86, ARM
• GPUs: NVIDIA CUDA & TensorRT
• NPUs: Rockchip NPUs (RV1109, RV1106, RK356x, RK3588)
• ANE: Apple Neural Engine (CoreML on macOS/iOS)
• Platforms: Linux, macOS, iOS, Android

Ready-to-Use SDKs

• Python package via PyPI: pip install inspireface
• Android SDK via JitPack
• Precompiled C/C++ libraries
• Docker-based multi-platform builds
• React Native module via JSI/Nitro Modules

Performance

On Apple devices using ANE (e.g., iPhone 13), the full pipeline of Face Detection + Alignment + Feature Extraction completes in <2ms, making InspireFace ideal for real-time applications.

Easy Integration

InspireFace is developer-friendly with bindings for:

• ✅ C/C++ (CAPI and C++ header interface)
• ✅ Python (ctypes interface and examples)
• ✅ Java / Android (JNI bindings)
• ✅ React Native (via react-native-nitro-inspire-face)

Quick Python Example:

import cv2
import inspireface as isf

session = isf.InspireFaceSession(isf.HF_ENABLE_NONE, isf.HF_DETECT_MODE_ALWAYS_DETECT)
image = cv2.imread("face.jpg")
faces = session.face_detection(image)
print(f"Detected {len(faces)} faces")

Commercial Support

Need help integrating InspireFace into your product? Looking for high-accuracy models or custom deployment support?

📧 Contact: contact@insightface.ai