Image Obscuration within Moir´e Patterns

Description

The pointwise multiplication of banding images – alternating black and white stripes, arranged in linear or sinusoidal patterns – often leads to interesting and unique visual effects. One family of combined banding images, referred to as Moir´e fringe patterns, is the result of interference between periodic bands. Moir´e images have two components: the base band image and the revealing bands. As the revealing mask is translated across a processed image, different amounts of shift will cause obvious visual effects, particularly beating animation. In this project, we investigate different geometric transforms and image modification techniques to obscure one or multiple images within a single Moir´e band pattern, both in black and white and in color.

Some image enhancement methods

Extracting bottleneck features for an image in tensorflow

https://web.stanford.edu/class/ee368/Project_Autumn_1617/Reports/report_spelman_molner.pdf