Face on Edge: Real Time Face Recognition on Snapdragon 845

Executive Summary

Face recognition is an algorithm/technique to identify faces by their names/IDs. A face recognition system captures an incoming image from a camera, crops and localizes the face, and extracts unique features from it. These features help build a classification model to recognize/classify faces. This paper’s purpose is to develop a face recognition system and deploy it on the Eragon 845 board, a Hardware Development Kit (HDK) featuring the Qualcomm® Snapdragon™ SD845 processor. The developed face recognition application is tested on Oneplus 6, an Android-based mobile device built.

on a Snapdragon™ SD845 processor. The classification model, Extreme Value Machine (EVM) , was recreated in Java Android, and the produced application’s performance was evaluated on the device’s

various runtimes (CPU, GPU, GPU_Float16, and DSP).

In this whitepaper we have discussed how we have developed and deployed the Real Time Face Recognition on the Eragon 845 and tested on the Oneplus 6 phone, both of which are based on the SD 845 SoC.

Project Highlights

  • Snapdragon Neural Processing Engine (SNPE)
  • Facial recognition pipeline
  • SSD MobileNet FaceNet
  • Extreme value machine Performance of face recognition pipeline
  • Facial recognition application demonstration
  • Application Area
To Download This Resource
Fill in the details below

Reference Designs

Our Work





Device Partnerships
Digital Partnerships
Quality Partnerships
Silicon Partnerships


Products & IPs