Done as a part of my Summer Internship in Madras Institute of Technology Campus, Anna University, Chennai This project deals with identification of different spiecies of plant by processing their leaves. The dataset is a collection of 4 available datasets namely Folio, Flavia, Swedish and 17Flower.
What things you need to install the software and how to install them
You will be needing Python 2.7 or higher.
The following python packages will be needed
1.Numpy
2.Scipy
3.Scikit-Learn
4.glob
5.OpenCV (cv2)
6.H5py
7.os
8.JSON
9.Datetime
10.Time
11.cPickle
12.Seaborn
13.Tensorflow
14.Keras in tensorflow backend
The above code uses keras package in tensorflow backend. To use it in the backend install tensorflow as er the instructions in the Keras with tensorflow website. Then navigate to your home and in the file .keras/keras.json add the code from the keras.json file in the repository. To use tensorflow it is recommended to use a virtual environment installation. To activate the virtual environment navigate to /home//tensorflow/venv/bin and then tyoe the command source activate
This program uses the pre-trained convolutional neural network model of VGG16 from the Keras package to extract the features. Then the extracted features are kept in the output folder in files called features.h5 and labels.h5
Then we use different classifing algorithms like K-nearest neighbour, Bagging, Logistic Regression etc to clssify the testing set.
The four datasets namely Folio , Flavia, 17Flower and Swedish can be downloaded from the above links