how to apply confusion matrix?

I have applied three methods i.e. chain code,pixel matching and morphological operations. Depending upon these methods three classes are there in which I have classified tool state i.e. normal, worn or unknown. If for example my actual output is 1- for normal, 8- worn,1- unknown and predicted are 2-normal,1- unknown and 7-worn. How it can be represented in confusion matrix to determine false positive,true positive etc.

Answer 1

[Image Analyst]

You can't just have bulk counts for how many are in each class - that's not sufficient. You need to know for each sample image, what the measured/predicted classification was, and what the actual/true classification was. Once you know that, it's trivial to construct the confusion matrix.

[messaoudi nada]

Dear @Image Analyst ,i have target= 640*1 double and output=640*1 double , i calculated confusion matrix by the matlab function confusionmat , when i tried to plot the confusion matrix by the function plotconfusion " plotconfusion(target ,output) , but it didn't work ,, can u help me , i dont know what's wrong in my code

[Image Analyst]

Attach your target and output in a .mat file.