Message Boards Message Boards

[WSS17] Citizen Science Project: Deeper Learning with Leaves


Dear All,

The goal of my project is to create a universal, interactive database for leaf identification while providing education opportunities on Computational Thinking and associated skills. The database itself is useful for research purposes and for hobbyists and other enthusiasts alike. The concept itself is an illustration about the possibilities that computational thinking can offer; a concrete human-machine interface that helps solving, or contributing to solve, a real world problem. For further technical details, please find the entire code here.

The outcome of the project is a web deployed user interface in which users can take a picture with a smartphone (or upload an image) of a leaf, with added possibility to identify the plant if already known (you can see an image of the user interface below, please, find more information on how to deploy it here):

Web form user interface created with CloudDeploy and FormFunction.

The output page is meant to provide links to more information related to the functionality of the web form together with information on how a user can contribute or learn more, such as by becoming an official verifier (direct link to the plant identification test will be provided) or by becoming a Wolfram Certified Instructor. The uploaded image is processed by a pre-trained neural network and the results are automatically displayed on the output page (you can find the code also here):

An output page with results from a neural network and additional information.

Simultaneously, a hyperlink with image information is emailed to a network of human verifiers, who have been pre-approved by passing a plant identification test:

A hyperlink with image information is automatically sent to pre-approved verifiers.

The verified images are automatically added to a Databin. The final result is a web deployable Human-Machine Interface which can also be easily adjusted for other use cases, such as other identification objects or by implementing other methods of identification. The web forms are created by using CloudDeploy with FormFunction. The data is stored in Databins and URLShorten Hyperlink is automatically sent to members of pre-created PermissionGroup.

Future work includes optimization of the neural network and plant identification test. In addition, the plan is to offer related workshops and create adaptive lesson plans for informal education, K12, college / university and professional training and provide the services also in other languages. The workshops would also include operation of Raspberry Pi and camera module, which would be used to take images of the leaves and upload them with the web form described above.

Thank you for reading for all those who made it this far! Any comments and suggestions are highly appreciated!

Happy wishes,

Daavid Väänänen

Every great and deep difficulty bears in itself its own solution.
It forces us to change our thinking in order to find it. — Niels Bohr

Keywords: education, citizen science, computational thinking, neural networks, machine learning, image recognition, web form, cloud deploy, cloud computing, hyperlink, databin

POSTED BY: Daavid Vaananen
11 months ago

Group Abstract Group Abstract