Inspiration
Our team wanted to address the Accurate Colour Detection challenge as we felt it was interesting yet also meaningful. We also decided to tackle the Best User Experience challenge as our app was heavily designed with accessibility at the forefront.
What it does
Our application is designed to extract and report the names of colours found in photos. It accomplishes this by using the K-Means algorithm to find the average colour of a region of pixels in the image and then compares it to a list of 843 colours, finding the one with the smallest Euclidean distance in RGB space.
How we built it
We built our app over the course of 36 hours in Android Studio using Java.
Challenges we ran into
We ran into several challenges in trying to extract pixel data from an image, ranging from out-of-bound errors to grabbing the wrong pixels.
Accomplishments that we're proud of
The accomplishment we're most proud of is our user interface, which manages to be high quality and polished while still being mostly compliant with a11y standards. Our user interface was specially designed to be friendly to individuals with colour blindness or partial sight by using large icons and text, and a colour-blind-friendly palette, as we believe these two groups to be the largest userbase of this application.
To further assist in being accessible to the visually impaired, our application uses speech synthesis to verbally say the name of the colour selected, allowing those who may struggle to read text on a phone screen to still use our application with ease.
We are also proud of our colour detection algorithm, which through testing has shown remarkable accuracy even in situations where other applications would struggle, such as extremely bright/dark colours or heavily textured objects.
What we learned
We learned a lot about developing an android app, how to develop an accessible application, and image processing using the k-means algorithm.
What's next for Hue Hunter
We plan to port our application to both iOS to allow even more people to use this software. We are also looking to add a live capture functionality similar to many other applications on the market, allowing a user to just point their phone at an object and automatically tell what colour it is.
We also would like to add full voice control to our application, to make it even more accessible. We truly believe that everyone, no matter who they are, should be able to use our app.
Finally, we'd love to continue to improve our algorithm. We would be interested in allowing more advanced users to configure the algorithm's parameters. We also believe we can further improve upon telling the difference between extremely light and dark colours by mathematically examining the differences between RGB components.
Log in or sign up for Devpost to join the conversation.