Textualize

Front Page
First Test Image
First Test Image Response
Second Test Image
Second Test Image Response

Inspiration

The project was inspired by combining many cool ideas about extracting similarities between art pieces or images. This application allows the user to discover the main colors in an image and find other images with similar tags. This application is also tailored to people who have a degree of visual impairment. The app uses Google's text2speech feature that allows the user to hear what colors are listed under the image as well as the description.

What it does

An online service that prompts the user to input an image's URL. This leads to an output of the most prominent colors in the image as well as an audio queue to alert the user.

How we built it

Used OpenCV for the main color detection algorithm and used flask to develop the web application. The infrastructure's code was written in python and the user interface was made with HTML and CSS.

Domain Registration and SSL setup

We used https://domain.com to register the required https://textualize.online domain. We used https://letsencrypt.org/ to create the SSL certificate and secure the website with HTTPS. The website runs on Ubuntu Linux with Nginx and Flask.

Challenges we ran into

Errors with the website's deployment, and trying to optimize the color detection algorithm.

Accomplishments that we're proud of

Optimizing the color extraction algorithm and attributing the extracted hex values to similar primary colors

What we learned

The skills required to deploy a fully functional web application

What's next for Textualize

Use Google Cloud's Vision API to generate the most dominant colors as well as classify objects in the image. Also creating a representative caption for the image using image2caption neural networks. Develop an Android application to use https://textualize.online from your mobile device.