1
     
  • COMPUTER VISION &
    DEEP LEARNING
  • COMPUTER VISION &
    DEEP LEARNING
3
2

MOBILE VISION

 

Mobile Vision is a term used to talk about Computer Vision algorithms written and optimized for a mobile environment, namely smartphones and tablets.

 

Benefits

Running the code locally, on a device, has many advantages, such as:
No need of an internet connection: the user can have access to Computer Vision features even when there’s no internet connection available
Short response time: using mobile vision algorithms it’s possible to trigger real-time solutions with instant responses, for an effective user experience
Scalability: the solutions implemented on mobile are scalable to an unlimited number of users, as there’s no need to use a cloud framework

Applications

Image Tracking: it’s used in augmented reality to know, at any given moment, the position of an object on the screen and makes it possible to display augmented reality content in a realistic way
Structure-from-motion: recreation of an environment or of a 3D object using a mobile camera
Image processing: processing of the video flow of a mobile device in local mode
Identification and classification: object and place classification

Know-how

Our best example of Mobile Vision is the Pikkart-AR SDK. Developed entirely with proprietary software, Pikkart-AR SDK incorporates a large selection of algorithms optimized for mobile devices, such as ARM NEON and SSE.

 

We use the smartphone’s CPU structure and its SIMD instruction sets (NEON or SSE), and we write fully optimized algorithms for both ARM and x86 devices.

 

We are working with various start-ups and established companies to bring their vision to life and build amazing computer vision applications for smartphones and AR headsets.

 

Last update: Giulia Fiorenza Campagna  26/03/2018