What are IRCODE’s Limitations?

IRCODE strives to be the best it can at image retrieval (and we’re the best out there) however, there are several use cases that we are still working on getting better at. Here are a few that will challenge our technology. 

  1. Single tone or gradient tone images
    • The IRCODE technology works off of recognizing shapes and forms referred to as technically as – Feature-based Image Matching. Although in the future we will get better at recognizing distinct images with few features and even single colors, at the moment we are unable to differentiate between the color blue by one artist vs the color blue by another artist.
  2. Extreme angles and partial occlusions
    • Although we are pretty stellar at recognizing images at angles and even partially obscured or altered, sometimes the image recognition will struggle at extreme angles or with occlusions of more than 50%. 
    • If you encounter issues recognizing images, please adjust your position and try again. 
  3. 3D Non-planar Images and Sculptures
    • Image recognition works for 3D artworks and objects as long as you scan them from approximately the angle at which they were registered.
    • We will soon be adding functionality to the IRCODE app to register a full 3D representation that we extract from a short clip recorded with the app. This allows registration of sculptures, architecture and interiors and eliminates restrictions on viewing angles.
  4. Shaky or blurry image capture
    • We do fairly well still recognizing images with slight blur but it can lead to mixed results. We recommend holding your phone as steady as possible when scanning an image for the best possible results. 
  5. Replications of famous images or paintings
    • By design IRCODE will recognize the original work rather than the copy or altered version. IE Mona Lisa Terminator image.
      • Unfortunately if you are an artist who loves taking other famous works and altering them we will have a hard time registering your work unless it differs significantly from the original. How much different – As this is a new technology we are still measuring this and will update this section as we come back with quantitative results.
Need help with something else?