Fruit-classification

This is general machine learning application for classifiying different fruits.

Classifiers used

Experimented with the following classifiers:

• SVM
• K-nearest neighbours

Features Used

The following features are used:

• GLCM
• CCV
• LTP

Prerequisites

• Matlab R2017a

Instructions

• Open the save_segment_images.m file and the change the train_folder to give the full path of the folder containing the training images.

• Run the save_segment_images.m file which will save all the segmented images in a separate folder that can be used later. (Note: You can change the value of k for k-means segmentation on line 14)

• In order to evaluate the features on segmented images run the EvaluateGLCMFeatures.m, EvaluateCCVFeatures.m and EvaluateLTPFeatures.m files.

• Run combineFeatures.m file to generate different combination of Image features.

• All the features data will be saved in Training_Data.mat that can be used later. (Note: You can use the data inside this mat file for detailed classification using the Classification Toolbox by importing different features data)

• Open plotData.m file and change the classifier to trainknn for knn classifier and trainsvm for svm classifier. Also, uncomment one line from 47-57 according to the classifier you wish to plot the data for.