Album Comparison

Inspiration

I always though the phrase "Hearing colors" was hilarious, so I thought what pictures would sound like if they were read like music sheets.

What it does

My program is charming and full of...quirks? It asks for an image input of an album cover, it then looks through a database to see which database album matches the best. Once it picks a good match, it plays musical notes that relate to the pixel colors of that album.

How we built it

I used the mean standard error and structural system method to see how different the images were. To do this, I had to loop through the database and convert all of those files to the same dimensions as the input image.

From there it creates a dictionary where the MSE's are ordered from best match to worst. It retrieves the best match and creates a list of rgb values for every pixel in the best match image. I converted the rgb values to ranges of colors with assigned numbers from 1-12. This is because I have 12 notes available in my database. From there I call the program to play the notes assigned to certain numbers at random increments, creating a wonky rhythm.

Challenges we ran into

At first, I wanted to use JAVA, however my team wanted to use Python. Not even a couple hours later my team never came back so I did this by myself with little knowledge of Python. That was really difficult to learn and commit to.

It was difficult figuring out the different syntaxes and conceptual theory for the image comparison section of this code. IT was hard to visualize a way to subtract the differences between images and then standardize the resulting error to compare with other images.

Accomplishments that we're proud of

I did this by myself, although it's ugly, I'm incredibly proud of myself for not quitting half way like I wanted to. I'm proud of being able to have a better grasps on how loops and packages like ktinker work. Oh! When I got the color array to convert to play music notes was the highlight of my night! I can't believe I did that, I recognize there are a lot of kinks and issues, but I'm still incredibly proud for finishing this event.

What we learned

I learned a lot about different python packages and ways to debug. I spent most of my time getting help debugging which helped me understand the different issues and ways to investigate errors. There were shortcuts I was taught that I was never taught in class! I don't understand why professors don't share such useful knowledge. I'm significantly more comfortable working with Pycharm after working on it for almost 24 hours.

What's next for Album Comparison

Honestly, I want to learn how to perfect panels so they're more aesthetic. I'm used to JavaScript and html, so I might try to move it to those platforms.

Built With

• python