Meet PerchEye-SDK: Our Lightweight, Open-Source Face Recognition Solution

Facial recognition is no longer just a futuristic feature. Today, it’s powering secure access, personalized experiences, and automation in industries ranging from fintech and healthcare to retail and smart homes.

But as we worked on projects requiring face recognition, we noticed that many existing SDKs were too heavy, too expensive, or too rigid for fast-moving product teams.

 

So we decided to build our own.

 

We’re excited to introduce PerchEye-SDK — an open-source, C++-based face recognition SDK created by Onix to give developers a lightweight, production-ready foundation for facial detection and matching.

 

Why We Built PerchEye-SDK

PerchEye-SDK was born from real-world needs. In several internal and client projects, we needed a face recognition tool that was:

 

• Fast enough for real-time use
• Lightweight enough for edge and desktop
• Easy to integrate, modify, and extend
• Open and license-friendly

Existing solutions often required restrictive licensing or bundled features that didn’t align with our product goals. Rather than compromising, we invested in a clean, modular SDK we could use — and share with the community.

 

What’s Inside PerchEye-SDK?

PerchEye-SDK combines the power of OpenCV and dlib in a clean C++ architecture, designed to be easy to read, modify, and extend.

 

Here’s a closer look at what it offers:

 

 

You can find the full codebase, along with setup instructions and sample embeddings, on GitHub.

 

Architecture Overview

PerchEye-SDK is built around a few key components:

 

• camera: captures and streams video input using OpenCV
• detection: locates faces and encodes them into embeddings
• recognition: matches faces by comparing vectors against known identities

This separation makes it easy to extend the SDK, whether you're integrating facial landmarks, emotion recognition, or even your own deep learning models.

 

Who Should Use It?

PerchEye-SDK is ideal for developers and companies building:

 

• Biometric login or identity verification flows
• Attendance or access control systems
• Retail analytics dashboards
• AR and computer vision-enabled apps
• Edge or desktop-based computer vision tools

Whether you're building a custom fintech KYC flow or experimenting with face tracking in AR, PerchEye-SDK provides a solid, hackable starting point.

 

What’s Next?

We're actively developing new features to extend the SDK’s functionality. Here’s what’s coming soon:

 

• Liveness detection for spoof prevention
• Facial landmark tracking
• Emotion recognition (mood and micro-expression analysis)
• Hardware acceleration support for embedded and mobile platforms

We’re also open to community contributions and feedback — if you have ideas, use cases, or improvements, we’d love to hear from you on GitHub.

 

Final Thoughts

At Onix, we’re passionate about building tools that are not just powerful — but practical and usable in real-world projects. PerchEye-SDK reflects that mindset: fast, clean, flexible, and open.

 

If you’re working on something that needs face recognition — give it a try. And if you’re looking for help building secure, scalable computer vision solutions, let’s talk.
 
Explore more Open Source by Onix