Lab4 Wiki

I used 10 categories each one contains 5 picture. the categories includes: Coffee, Dance, Drinks,Fruit, Kebab,Music, Omelet, Pizza, Rice, Spaghetti

Explanations:

firstly it extract key descriptor via SIFT method, then clustering them with K-Means algorithm. the center of our cluster will be used for visual words, Then create a histogram based on these words, after that, train the image clasifier, here we used random forest, with these words, These steps will be done for each image to make the training data,

Finally test the accuracy of our model with the images in test folder,

Source code:Sourcecode

image classification

Input: Training Data: 75% Testing Data: 25%

model:****

Accuracy and Confusion Matrix: Accuracy