Portfolio Project

Sheet Music Watermark Removal & Upscale

UNet & VDSR Pipeline

Machine Learning Python Computer Vision
Preview image for Sheet Music Watermark Removal & Upscale

Context

I wanted the ability to participate in playing music as my church but did not have any sheet music.

Approach

  • Trained a UNet model on more 20,000 paired pages for watermark removal.
  • Upscaled 612×792 scans to 1700×2200 with Very Deep Super-Resolution (VDSR).
  • Wrapped functionality in a GUI for a simple pipeline.

Impact

  • Delivers cleaned and legible sheet music in <10 seconds.

Links