Skip to content

Abhishek-Nalawade/PbLite-Boundary-Detection

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 

Repository files navigation

PbLite Boundary Detection

License: MIT

About

Boundary detection is an important, well-studied computer vision problem. A simple method to find boundaries is to look for intensity discontinuities in the image, also known of edges. Classical edge detection algorithms, including the Canny and Sobel baselines, look for these intensity discontinuities. The more recent pb (probability of boundary) boundary detection algorithm significantly outperforms these classical methods by considering texture and color discontinuities in addition to intensity discontinuities. Qualitatively, much of this performance jump comes from the ability of the pb algorithm to suppress false positives that the classical methods produce in textured regions. This project is a simplified version of pb, which finds boundaries by examining brightness, color, and texture information across multiple scales (different sizes of objects/image). The output of the algorithm is a per-pixel probability of boundary which outperforms the Canny and Sobel Edge Detectors.
This project implements a simpler version of the algorithm presented in this paper

Details

Input image

Derivative of Gaussian Filter (DoG), Leung-Malik Filter (LM) and Gabor Filter Banks:

Texton, Brightness and Color Gradient maps:

Results:

Sobel, Canny, and PbLite edge detections

Libraries Required:

  1. Numpy
  2. OpenCV
  3. Matplotlib
  4. Scikit-Learn