1 of 10

How Industries Survive?

EVOLUTION

  • Entertainment Industry : From tv to Web
  • Internet Industry: From web to App
  • Manufacturing Industry: Manual workers to Automation
  • Banks: Gone Digital
  • Retail: moved to E-commerce
  • Travel and Hotel Booking: Gone digital

In last 10 years

2 of 10

What about Security and Health Industry?

  • For Security it was Computer vision
  • For Eyewear Industry it was Augmented Reality
  • For Healthcare Industry it was Online consulting and record keeping

Next is What?

3 of 10

EYE ASSIST and 3D facial Construction

4 of 10

Calculate Power of Lens required for correct vision using Laptop/ Smartphone without any external help or Machine

and construct 3d Facial image for applications of Augmented Reality.

5 of 10

Traditional Methods:

  • Snellen Chart
  • Auto-refractometer

6 of 10

How to use Eye Assist?

  • Look towards the camera of the Laptop/Mobile
  • Read the text being shown on Instructions Page
  • Move away from Camera until you can read the text carefully
  • Stabilize at that Position
  • Voila it’s done

​

How to use 3D Facial Constructor?

  • Just look at the WebCam and press the interrupt key
  • That’s it, it’s done
  • Further the whole pipeline is automated, the 3D face object is uploaded to Azure Cloud and is then consumed by the Android App

​

7 of 10

3D Facial Construction

​

When the Position is stabilized a frame is captured and 3D model is constructed out of it, which can be used for any Augmented Reality based Application

​

​

8 of 10

How it Works?

  • A Convolutional Neural Network(CNN) trained on i-Bug 300-W facial Landmarks dataset to get Facial Landmark co-ordinate points on Face.
  • Co-ordinates of points representing both eyes are calculated.
  • From which Cornea to Cornea distance is calculated.
  • Another Pre-Aligned model Calculates distance of eye from Camera at which the person can see clearly.
  • Now Depth perception(that too from a Single Camera) is used to Calculate power of Lens the person require.
  • 3D - Facial construction is done using Position Map Regression Networks
  • 3D objects are then sent to the Azure Cloud, which is then used by the App for Augmented Reality based Application

9 of 10

How it helps?

  • 3D facial reconstruction can be used to construct images of Criminals and it can be known how they actually look and many permutations of how they would look with different facial aspects like moustache, beard and long hairs, etc.
  • Prediction of facial paralysis
  • Makes use of Auto-refractometer and Snell’s test redundant
  • Ease of performing the test
  • 3D- face model can be used for AR- based applications like trying Spectacles at home at much better measurements and higher accuracy
  • E-commerce sites like Lenskart can integrate it with AR service and directly deliver the spectacles
  • Free of Cost
  • No internet dependency

​

10 of 10