matlab - How to implement the conditional probability for each pixel? -
i have asked 'very broad' question yesterday link build conditional random field based energy function images. got negative feedback comments , think should modify question , make more specific.
here have got bunch of images cow on grass , sky @ background. want segment cow grass , sky (a toy problem only).
i firstly over-segment images using super-pixel method, , have got ground truth of labels of 10 training images. here example,
then pass these super-pixel patches filter texture features , saved them feature vectors cow (c), grass (g) , sky (s)
my question how using matlab implement conditional probabilities of these 3 classes?
like p(x_i|c_c), p(x_i|c_g) , p(x_i|c_s)? x_i each super-pixel, c_x 3 classes.
i think webpage mentioned using matlab hist function. not sure how , why that. please give me basic applicable hints not complicated papers. lot. a.
Comments
Post a Comment