The project includes 2 files:
1. features.py
2. classifier.py

training dataset - http://flavia.sourceforge.net/


-----features.py extracts the features related to shape, color, and texture

function prototype

def summation(image)
def lacunarity(image,p)
def skewness(image,avg,stddev)
def kurtosis(image,avg,stddev)
def create_dataset()

lacunarity(),skewness(),kurtosis(),summation() are helper function for the named features
create_dataset() is to extract all the features.
The created dataset is put in the csv file.

-----classifier.py classify the leaves into one of the 32 species

functions prototye

def preprocess()
def PNN(X_train, X_test, y_train, y_test,X_dummy)
def SVM(X_train, X_test, y_train, y_test,X_dummy)
def ANN(X_train, X_test, y_train, y_test,X_dummy)
def dummy_data()

names of the prototye are self-explanatory

-------environment used:
  OS Used: Windows 10 (64 bit)
  Language: python 3.5
  Environment: conda 4.5.11
  libraries:
    OS  
    string
    numpy
    pandas
    scikit-learn
    neupy
    cv2
    mahotas

-----To classify run command on cmd:
    1. python features.py to create the dataset.
    2. python classifier.py
    3. Input the name of the classifier you want to train,
    enter either ANN or SVM or PNN (input is case sensitive)
    4. results of the classification on flavia and dummy dataset 
       will be shown