Image Segmentation

Marks: $15$

Q$6.$ Thresholding is a basic step to convert a gray scaled image into a binary image using a specific threshold value. To find this threshold value is the most critical part. Assume that we have designed a basic algorithm to calculate adaptive threshold $T.$ Algorithm for computing $T$ is

Use initial threshold to put input image gray levels into two bins $($classes$)$

Compute Global mean $(M_G)$ for whole image, local means $(m_1$ and $m_2)$ and scatters and $($:$S_2\}$ for both bins

Compute between class mean $(M_B)$ and scatter $(S_B)$ using $\frac{?}{abs}(m_1-m_2)$ and abs $(S_1-S_2)$

Compute new $T$ using $T=M_B+S_B$

Repeat steps $(1-4)$ to find such $T$ which maximizes the product of $M_B$ and $S_B$

You are given a $55$ image in fig$\mathrm{.}3.$ Assume initial $T=100.$ Run one iteration of above mentioned algorithm to find $T$ and draw binary image after first iteration. Show all steps of your processing.

$\backslash$table$[[10,70,110,78,115],[210,45,51,51,65],[209,245,187,71,35],[10,195,205,199,201],[10,$Figure$-3,55,65,]]$