"No one is harder on a talented person than the person themselves" - Linda Wilkinson ; "Trust your guts and don't follow the herd" ; "Validate direction not destination" ;

June 23, 2022

Retail & Vision - 2.0 Learning

Interesting paper and discussions link

Session #1 - AI for Retail Problems

  • AI in retail

  • Same product with variations

  • Leverage a combination of techniques



  • Approach to learn


Session - #2 - Vision for Fashion

  • Customizing existing designs to make it better

  • Fashion++ alter the design

  • Training based on 
  • Classifier for fashion estimation
  • Suggesting changes


Paper - C O M P U T E R V I S I O N A N D D E E P L E A R N I N G F O R R E TA I L S T O R E

M A N A G E M E N T

  • Automatic recognition of items on store shelves and obtaining a reliable 3D reconstruction of an environment for navigation purpose
  • Planogram - This arrangement is carefully planned to maximize sales and keep customers happy, currently, however, verifying compliance of real shelves


  • Planogram as a grid-like fully connected graph
  • Planogram compliance issue related to the checked node is reported (i.e., missing/misplaced item).
  • Camera would locally run only model relevant for the Aisle
  • Given a shelf image, we perform first a class-agnostic object detection to extract region proposals enclosing the individual product items

  • we propose to deploy an image-to-image translation GAN together with the embedding CNN and to optimize the whole architecture end-to-end

More Reads

Keep Exploring!!!

No comments: