matlab code for entropy based local binary pattern

can anyone send me the link for entropy based local binary pattern(ELBP) .. thanks in advance

[Imran Riaz]

You found the code for ELBP? If yes plz share with me also.

Answer 1

[TED MOSBY]

Hi Annie,

To get started you can follow the below steps:

• Preprocess your image by converting to grayscale and then perforn normalization.
• Use MATLAB’s built-in 'entropyfilt' function to compute the entropy for each pixel within a local window (of a specified size). This filter computes the entropy within a neighborhood around each pixel.
• After computing the entropy, normalize the result to the range [0,1]. This standardization helps when comparing the center pixel’s entropy with that of its neighbors.
• For every pixel (except for the borders, to avoid boundary issues), sample the neighbor pixels in a circular pattern around the center pixel. The number of neighbors and the radius of the circle can be adjusted.
• Since neighbor coordinates may not fall exactly on pixel locations, bilinear interpolation ('interp2') can be used to estimate the entropy value at these points.
• For each neighbor, if the neighbor’s entropy is greater than or equal to the center pixel’s entropy, assign a binary value of 1; otherwise, it is 0. This forms a binary pattern for each pixel.
• The binary pattern can be converted into a decimal number. This number represents the LBP code for that pixel.
• Then a histogram can be created by counting the frequency of each LBP code over the valid (non-border) region of the image.
• Finally, you can calculate the entropy of the LBP histogram. This overall histogram entropy will serve as a global texture descriptor for the image.

Here is the documentation link for 'entropyfilt' :

https://www.mathworks.com/help/images/ref/entropyfilt.html?searchHighlight=entropyfilt&s_tid=srchtitle_support_results_1_entropyfilt

Hope this helps!