Project Keywords:
    | MATLAB
    | Biometrics
    | Java Application
    | Data Processing
    | Data Classification

Implementation Process:
This biometric project has five main components: (1.) data collection, (2.) feature extraction, (3.) database training,
(4.) classification, and (5.) a graphical user-interface application.
  1. For data collection, we took over 300 images of leaves from more than 20 different plant species. The images were
       taken with a white background and were of a close-up and front-facing leaf.
  2. Once the data had been collected we extracted seven unique features to use for classification. These features
       included the area of a leaf's skeleton, leaf surface area, physiological width, and leaf perimeter. The feature extraction was
       done using MATLAB.
  3. Next we created a classifier using the Kth Nearest Neighbor algorithm. This classifer takes an unknown leaf's features
       and compares it with our database (step #4) to recognize which tree the leaf belongs to.
  4. After the classifier had been built, we ran our feature extraction algorithm on our collected images generating our dataset.
       To train our application, we trained our classifier to maintain a logged database of the collected data.
  5. Lastly, we created a Java application which accepts an uploaded leaf image and runs the feature extraction algorithm.
       After this it then runs the extracted data through the classifier to output the resulting tree species and
       image the leaf is from.

Contributers: Disha Chaubey