On-Premises Facial Recognition Server

We are a Singapore technology startup with a focus in creating smart vision solutions at the edge.

We have created a technology that enables developers and system integrators to create facial recognition applications with state-of-the-art recognition accuracy of 99.73% on the Labelled Faces in the Wild (LFW) benchmark.

From aged photos to poorly illuminated captures, our technology offers you options to deal with any kind of challenging scenario. Unlike traditional facial recognition technology, we use the very best, state-of-the art, deep learning models to deliver you best-in-class performance.

A simple REST JSON API call is all that is needed, to enable a developer or system integrator to create applications tapping on our facial recognition technology. The API can be called using any programming language or framework of choice.

Accuracy

• LFW Benchmark
• Fast model: 99.62%
• Accurate model: 99.73%

NIST IARPA IJB-A

• FNMR @FMR=0.01, SSAI @ 0.027

NIST IARPA IJB-C

• FNMR @FMR=0.01, SSAI @ 0.022

Our technology is economically priced using low-cost hardware. The system requirements are as below:

System Requirements (Physical machine)

• OS: Ubuntu 18.04/20.04 LTS Desktop Server
• CPU: Intel Core i7 8th gen or later
• Clock Speed: 3.2GHz, 6 cores, 12 threads
• RAM: 8GB
• Free Disk Space: 1GB 
• Registration Speed: 4-5 Requests per second ( is the speed of registering a person into the facial recognition database)

Search Speed

12-13 RPS (is the speed of searching  for a person in the database with one image)

* In addition to its initial size, the product consumes 106.5MB per 10,000 enrolled persons with 1 picture each.

You will be able to:

• Integrate this technology with your biometric access control for a non-contact flow.
• Combine cloud deployments with our edge-enabled technology solution for seamless operation even when the network is down.
• Integrate our solution with existing attendance systems.
• Develop unique solutions using facial recognition.
• Integrate our technology in mobile banking Android apps for face verification.
• Create digital IDs to enable safe and secured access.

Our technology supports edge AI deployment and is priced economically, using commodity hardware such as CPUs, mCPUs.

All data is stored locally on your machine thus ensuring the privacy of the data and our API can be easily integrated into existing products like building management, access control and attendance systems.