Skip to content

Shadow removal by subregion matching and illumination transfer

Notifications You must be signed in to change notification settings

cdbharath/shadow_detection_and_removal

Repository files navigation

Digital Image processing - NITT course project

Shadow detection and removal

Reimplementation of the paper Efficient Shadow Removal Using Subregion Matching Illumination Transfer with a slight twist

Implemented an unsupervised segmentation algorithm employing autoencoders for detection of shadow regions. Gabor filter is designed to identify the texture features in the images. Illuminance transfer techniques are deployed to remove the shadow regions with the help of the acquired textures. Implemented the algorithm on SBU Shadow Datasets obtaining good results. More work is needed on the boundary processing of the shadows to improve the accuracy.

gesture detection

The flow of the shadow detection and removal algorithm

Reference:

  1. Efficient Shadow Removal Using Subregion Matching Illumination Transfer
  2. Unsupervised Segmentation

About

Shadow removal by subregion matching and illumination transfer

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages