Mood Color

• 

  Half done typing

• 

  Animation finished

• 

  Paragraph completed, animation on the way

• 

  Sad face produced from sentiment analysis score

Inspiration

Partly motivated by the machine learning API and not knowing what to do, we decided to go ahead with our first idea. We want to make an interactive image based on the content of text, which is different from traditional static style. This project may extend to the field of gaming, web design etc.

What it does

Our program analyses the sentiment of user entered text, then changes the colour(hue) of an image according to the results that we got from the analyses. The analysis is sentence based, and for each sentence, a mood is generated and reflected in the picture. The picture at bottom right shows how mood changes from sentence to sentence, and the effect on the picture can be observed on the top right picture viewer.

How I built it

It is built using python, ah, based on pygame, also with help from OpenCV and the machine learning API from indico.

Challenges I ran into

Image processing was much harder than we expected. Pygame doesn't really provide an impressive UI.

Accomplishments that I'm proud of

We re-invented many wheels!

What I learned

How to bell-curve mark (when we were trying to distort image). Linux is good.

What's next for Mood Color

UI: a more user friendly UI Current functionalities: For now, the picture shows its overall changes of the text periodically. A better functionality can be achieved by only presenting the result from current text (need more time to implement). More functionalities: Many many more features can be supported. More image changes can be implemented: such as rotating, blurring, twisting, cutting and pasting images, change the colour balance of the pictures etc. In addition, face recognition can be supported to make the person funny depending on different texts

Built With

• python