In Graphical Translation

Senior thesis for Computing and the Arts by Felicia Chang

Running the App

Install dependencies (preferably to a virtual environment)

pip3 install -r requirements.txt

Run the app

python app.py

Production

You can find the live app at https://felicia-senior-thesis.herokuapp.com/

Design

Image as Type

Frontend

You can find the html file for Image as Type under templates in p5js.html. The html file merely provides the skeleton. The graphics are created in sketch.js and createWord.js found in the scripts folder.

Backend

Image processing is done in app.py using the functions getImage and getCircles.

Type as Image

Frontend

You can find the html file for Type as Image under templates in depth.html. Like Image as Type, the graphics are processed in modelDepth.js found in the scripts folder.

Backend

Image processing is also done in app.py using the function getDepth.

Resources

The Monodepth Library (https://github.com/nianticlabs/monodepth2) for the depth map algorithms. WebGl Fundamentals (https://webglfundamentals.org/) for providing guidance in navigating the WebGL library. OpenCV (https://opencv.org/) for image processing API's.

Acknowledgments

Thank you to Holy Rushmeier for technical guidance and Julian Bittiner for artistic guidance.