The Verification Process

There are five core steps in the verification process. Each step includes advanced imaging techniques and neural network algorithms. 

 

1) Image Capture 

• Dynamically control camera to rind & track eyes, set focus and exposure
• Capture images
• Assess Image Quality

  

2) Segmentation

• Segment images to find region of interest (ROI)
• There are 4 ROIs to segment (left and right side of each eye
• Leverage image processing and neural network algorithms such as Color Clustering and Active Contour           

           

3) Image Enhancment

Enhance the images to maximize the vascularity features. 

• Exposure and focus image enhancement
• Vein enhancement 
• Color layering separation
• Vascularity measure

4) Feature Extraction 

Based on quality, vascularity and other parameters, extract relevant features

• Texture based method
• Landmark based method
• Sparse codes method
• Feature fusion

5) Matching

 Using various methods, match to template and fuse scores across images

• Distance metric based
• Neural-Network
• Score level fusion
• Robustness analysis