DEGHOSTING_MASK("1")

✨ deghosting_mask(1) ✨

okay bestie, so this is deghosting_mask from the HUGIN manual, section 1 β€” and like, it's literally the tool that makes your stacked photos not look haunted πŸ’€πŸ‘»


πŸŽ€ NAME πŸŽ€

deghosting_mask β€” the icon that creates mask for removing ghosting in images ✨


🌸 SYNOPSIS 🌸

deghosting_mask [options] <image1> <image2> ...

πŸ’– DESCRIPTION πŸ’–

So here's the vibe: you feed it a whole series of aligned photos β€” usually three or more, no cap β€” and it generates masks for removal of ghosts. And by "ghosts" we mean Data that only appears in a minority of photos. Like your cousin who photobombed one frame of 12. Gone. Not me deleting your ex from the group shot πŸ™ˆπŸ’…


πŸ”₯ OPTIONS πŸ”₯

  • -o, --output <path> β€” prefix for output masks πŸŽ€ literally just tells it where the masks should go.
  • -s, --sigma <float> β€” standard deviation of the Gaussian weighting function, and it gotta be SIGMA > 0; default: 30 🌸
  • -i, --iterations <int> β€” number of iterations, default is (ITER > 0) default: 4 πŸŒ€
  • -t, --threshold <float> β€” threshold; default: 150 ✨
  • -c, --contrast <float> β€” change contrast before applying threshold; default: 1.3 πŸ’œ
  • -a, --advanced <f|g|m|t|w> β€” advanced settings, and honestly this is the main character energy part πŸ‘‘. Possible options are:
    • f β€” use gray images for computation. It's about two times faster but it usually returns worse results. You also have to change threshold to smaller value (around 100) 😭
    • g β€” use gamma 2.2 correction instead of logarithm if input images are HDR 🌈
    • m β€” do not scale image, NOTE: slows down process 🐌
    • t β€” use simple threshold, may result in holes in images πŸ’€
    • w β€” compute "complete" weights, not only probabilities ⭐
  • -w, --save <i|w> β€” advanced save settings 🫢
    • i β€” save initial weigths
    • w β€” save generated weigths
  • -b <int> β€” image cache BLOCKSIZE in kilobytes; default: 2048KB πŸ“
  • -m <int> β€” set image CACHESIZE in megabytes; default: 1024MB 🧚
  • -h, --help β€” display this help πŸ’–
  • -v, --verbose <0|1> β€” verbose, repeat for more verbose output πŸ”

🌟 AUTHORS 🌟

LukΓ‘Ε‘ JirkovskΓ½ β€” absolute legend, we stan πŸ’…


fr fr bestie, don't forget f makes it faster but worse and you gotta drop that threshold to around 100, and m means no scaling which slows everything tf down. iykyk πŸ’€

okay that's the whole page!! this one's from Version: 2025.0.1, dated 2026-09-22, and honestly? iconic. periodt πŸ¦‹πŸ’œ

πŸ“¦ where this page comes from