Example-based image manipulation

Erik Reinhard
Max Planck Institute for Informatics (Germany)
Download paper

Play (33min)

Download: MP4 | MP3

Many color-related image adjustments can be conveniently executed by exposing at most a small number of parameters to the user. Examples are tone reproduction, contrast enhancements, gamma correction, and white balancing. Others require manual touch-ups, applied by means of brush strokes. More recently, a new class of algorithms has emerged, which transfers specific image attributes from one or more example images to a target. These attributes do not have to be well-defined and concepts that are difficult to quantify with a small set of parameters, such as the “mood” of an image, can be instilled upon a target image simply through the mechanism of selecting appropriate examples. This makes example-based image manipulation a particularly suitable paradigm in creative applications, but also finds uses in more technical tasks such as stereo pair correction, video compression, image colorization, panorama stitching and creating night-time images out of day-light shots.

You may also like:

  1. Color Image Super Resolution: A Two-Step Approach Based on Geometric Grouplets
  2. Spectral Image Prediction of Color Halftone Prints Based on Neugebauer Modified Spectral Reflection Image Model
  3. Accurate appearance-based visualization of car paints
  4. The impact of image-difference features on perceived image differences
  5. Colour based image retrieval with embedded chromatic contrast

  • Share
1 Star2 Stars3 Stars4 Stars5 Stars (No Ratings Yet)
Loading ... Loading ...