Fast Fourier Transforms in Python

• 

  output

• 

  input

Inspiration

My father building his first computer application based on Fourier image analysis.

What it does

Generates a 3d plot of frequencies (colored in heat map colors, high frequencies are red and low frequencies are black).

How I built it

Used online examples and lots of research papers to understand FFT

Challenges I ran into

Matching the axes correctly so that colors would show up based on frequency instead of y coordinate of the image. Getting image filtering to work Getting sound files to work Getting enough photos for fft comparison to be effective

Accomplishments that I'm proud of

Getting something visually appealing for the 3d plot that doesn't take 15 minutes to generate

What I learned

I can't think under time pressure; I won't be satisfied easily.

What's next for Fast Fourier Transforms in Python

Using the plots of the fourier transformed images to find the average frequency of the image/sound file, then comparing with other images/sound files to find trends. (Attempted; data size too small to gather any conclusions) Using similarity of frequency of topographical maps in certain cities to understand how cities are made. (Attempted; couldn't find topographical maps without any markings on them) Modifying FFT values and inverting the result to get a product with less "noise" (visual/audio) (Attempted; somewhat completed) Analyzing TV/video to see trends in frequencies on screen (Haven't attempted) Predict outcomes using a small amount of data (ex. location logs of a driver --> prediction of location)

Built With

• python