Find People in Event Photos Instantly: How 3DiVi Face SDK Transformed a Photography App

Finding the right person in thousands of event photos is a task that can take hours. Our partner, a photography software developer, recognized this challenge and aimed to improve their desktop application to help photographers organize and sort large image collections more efficiently.

Their Windows-based tool, built using C++, was already a powerful solution for cataloging and selling photos online. But as photo libraries grow, users need a faster way to find their images. That’s when they saw an opportunity—integrating face recognition to speed up image sorting and retrieval.

Our partner integrated 3DiVi Face SDK into their C++ application, enabling the following intelligent features:

• Face Recognition: Users could quickly search for and find all images of a specific person in the collection, or check if someone appears within a selected group of images.

• Eye Openness Estimation: This feature helped detect images where people have their eyes closed, ensuring only the best shots are included in the final selection.

The partner faced no conceptual challenges with the integration, as 3DiVi Face SDK natively supports C++ integration. In fact, it’s our preferred API since the SDK code itself is written in C/C++.

For the app's architecture, we adopted a simple yet effective approach, following the core Face Recognition pipeline:

To strike the perfect balance between speed and accuracy, a few clever strategies were employed. The goal was twofold: identify as many faces as possible in the photo while ensuring the process was lightning-fast. Here’s how we achieved it:

• Detector Cascade: If the primary, balanced detector fails to find any faces in the image, we trigger a different model — one that’s better suited for specific scenarios, such as portraits or crowded settings. This layered approach ensures we never miss a face.

• Thumbnails for Faster Processing: When an image is identified as a portrait, we can speed up detection by processing a smaller preview instead of the full-resolution image. Face detectors handle smaller images much faster, but this results in a less detailed biometric template. To minimize this effect, we only use this method for images specifically marked as portraits.

Better User Experience
The new features completely transformed the photographer’s workflow. Photographers no longer waste time filtering out photos with closed eyes, and can easily gather personal photo collections for each person at an event.

Instant Image Search
The right images can now be found in seconds, even within large photo collections.

Easy Licensing Model
The flexible 3DiVi Face SDK licensing model allowed unlimited deployments, making it easy for the partner to scale their solution without restrictions.

With 3DiVi Face SDK, the partner turned their desktop software into a smart, AI-powered tool that saves time and improves the workflow for professional photographers. The solution was successfully integrated and tested within one month.

Looking to upgrade your software with top NIST-rated face biometrics? Reach out for a free consultation to see how 3DiVi Face SDK can drive your business forward!