TECHNOLOGY

Orpix combines state of the art computer vision and deep learning technology with its scalable web-based platform to bring a unique capability to its customers.

COMPUTER VISION

The Orpix team has years of experience in the field of computer vision. Led by technical advisor Professor Serge Belongie, Orpix has developed a variety of image recognition algorithms which combine both traditional image matching with state of the art deep learning. This combination allows the algorithms to complement each other in order to maximize performance and accuracy.

Scalability on The Cloud

Orpix solutions are deployed on the cloud to take full advantage of the cloud’s flexibility and scalability. The solutions and cloud resource management are both provided via a simple to use web interface. This allows customers to allocate resources as needed, and to start/stop the server at a schedule that they choose.

The combination of our unique computer vision technology and scalable web-based platform on the cloud allows for the rapid development of accurate and robust solutions to enhance your business process.

Orpix Solutions are built on top of a unified platform that can be summarized in the following steps:

System Setup

Create a domain – A domain defines all of the data needed to create and deploy a detector

Define meta-data – The attributes that each category in the domain will have

Define categories – Create the categories required for detection and/or manual labeling

System Setup

Create a domain – A domain defines all of the data needed to create and deploy a detector

Define meta-data – The attributes that each category in the domain will have

Define categories – Create the categories required for detection and/or manual labeling

Data Collection

Capture training images – Gather zip files and/or video files that will be used for training/testing

Upload images – Upload and extract image/video data into the labeling solution

Label images – Label a training set quickly using both automatic detection and manual labeling in an easy to use web interface

Data Collection

Capture training images – Gather zip files and/or video files that will be used for training/testing

Upload images – Upload and extract image/video data into the labeling solution

Label images – Label a training set quickly using both automatic detection and manual labeling in an easy to use web interface

Training

Setup train parameters – Adjust parameters that control speed and accuracy

Train – Launch a training job

Test – Evaluate the accuracy of the trained detector

Training

Setup train parameters – Adjust parameters that control speed and accuracy

Train – Launch a training job

Test – Evaluate the accuracy of the trained detector

Deployment

Select data center – There is the option to deploy the service on any server that meets requirements, but it is optimized to be deployed on AWS

Configure cluster – Configure the cluster for the number of machines you want to run the service on

REST API integration – integrate the solution into your application using the Orpix REST API

Deployment

Select data center – There is the option to deploy the service on any server that meets requirements, but it is optimized to be deployed on AWS

Configure cluster – Configure the cluster for the number of machines you want to run the service on

REST API integration – integrate the solution into your application using the Orpix REST API

Interested in learning more?