I’m studying the LBP algorithm and reading the paper Face Detection and Verification using Local Binary Patterns, Y Rodriguez which is a PHD thesis paper. On the page 21 (section 2.2.2 weak classifiers), the author mentioned that:
A weak classifier hp (x) consists of a look-up table of 2^9 − 1 = 511 bins
I cannot understand why there are 2^9-1 bins. What I can understand till now is that the LBP code for a specific pixel is 8 numbers either 0 or 1 calculated from its 8 surrounding pixels. However why the look-up table contains 2^9-1 bins? Any help is appreciated!
I suggest you to read the paper “Multiresolution Gray-Scale and Rotation Invariant Texture Classification with Local Binary Patterns” written byTimo Ojala, Matti PietikaÈ inen, Senior Member, IEEE, and Topi MaÈenpaÈa.
You will find the answer in chapter 2.2
For example, the pixel is surrounded by 8 pixels, the lbp bin will be 256(-1) = 1111 1111,by the same token, surrounded by 9 pixels bin 512(-1) = 1111 1111 1…you can think the pixel is surrounded by a circular chain formed by 9 pixel.