Skip to content

Brmanzo/dsp-tools

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 

Repository files navigation

DSP Tools

Welcome to DSP Tools! This project is a collection of image processing utilities, I created to practice programming algorithms and app development.

Features

  • Background Removal for PNGs and GIFs using:
    • Flood fill
    • Blob fill
    • Erosion
    • Aliasing/feathering
    • Cropping
  • Spatial Filtering
    • Blurring
    • Sharpening
    • Edge Detection
  • Extended functionality from preprocessing.py (imported from my clash_star_tracker project) to explore additional image processing applications.

Goals

I plan to add more useful features and tools in the future. Contributions and suggestions are welcome!

Future Ideas

  • File conversion
  • Pixel art downscaling
  • Image slicing/ Grid detection
  • FFT-based edge detection
  • Preprocessing tools for ML inference

How to use

python -m tools <tool> <input image> <args>

Tools:

  • remove_bg (.gif supported)
  • blur
  • sharpen
  • edge
  • preprocess_text

Options:

  • lakes (Remove internal white space if true)
  • no_alias (disable aliasing)
  • square_kernel (use a 3x3 kernel insted of diamond)
  • no_crop (To preserve the original image dimensions)

Modifiers:

  • flood_tol (Difference in lightness to flood over when removing background)
  • erosion_tol (Difference in lightness to erode when aliasing)
  • crop_slack (How much additional margin in pixels to leave when cropping)
  • crop_tol (Difference in lightness to detect background from subject)
  • fade_factor (controls how quickly the alpha fades (higher = quicker))
  • halo_pct (controls how many of the eroded layers to use for aliasing (0%-100%))
  • blob_tol (How large a blob must be to be removed (blob area/image area))
  • kernel_size (kernel size, 3 -> 3x3, 5 -> 5x5)

About

Developing image processing tools for practice and fun.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages