File Exchange

image thumbnail

Center-symmetric local binary patterns (CSLBP)

version 1.2 (1.25 KB) by

An extension of Local Binary Pattern (LBP), that is quantized in nature. Used as descriptor.

3 Ratings



View License

This computes the Histogram of CSLBP operator, which is an extension of LBP. Using CSLBP is really efficient for illumination and blur type of image transformation.

Comments and Ratings (7)

Mano R

Mano R (view profile)

Hi Junaid,
Can you provide me a code to use it like a SIFT descriptor. because i m using SIFT for feature extraction. can i use both SIFT+LBP for object recognition? can you help me out


Junaid (view profile)

Thanks Rouzbeh for your comments and suggestion. I will add parameters in function like T. I didn't use interpolation because in given code maximum 8 neighbors can be selected given radius 1. If I extend my code for multiple radius for neighbors then definitely I will add interpolation for neighbor estimation like LBP.

Thanks for comments.


You may also compute the value of the neighbors not located at the center of pixels by interpolation (i.e., b and d in your code) like the original LBP at:


You may also add the parameters as mentioned in the paper (e.g., T and N).


Junaid (view profile)

Hi Wilfre,
Thanks for your comment, CSLBP can be used in many applications. In this file I only gave how to compute histogram of CSLBP operator.

For some application it can be used as CSLBP based image for image detection or recognition, or it can be used as region descriptor. So far I have used it as as region descriptor like we use SIFT.

To use it like descriptor, then we divide the given region into blocks ( usually 4 x 4) and compute the histogram on each cell and finally concatenate them as single vector.

If you want an example as descriptor then I can share that code as well.


Hi Junaid,
Can you give an example application by image


Aravin (view profile)



Change the comments only.

MATLAB Release
MATLAB 7.8 (R2009a)

Download apps, toolboxes, and other File Exchange content using Add-On Explorer in MATLAB.

» Watch video