TEND-MFIT(1)

💖 TEND-MFIT(1) 💖

okay bestie, welcome to tend-mfit — it's a User Command, it lives in section 1 of the manual, and it is literally the moment you finally figure out what your DW images are actually telling you. 🧚✨

✨ NAME ✨

tend-mfit — estimate models from a set of DW images 🫶

🌸 SYNOPSIS 🌸

tend mfit [-v <verbose>] -m <model> [-ns <# starts>] [-ml] [-sigma
<sigma>] [-eps <eps>] [-mini <min iters>] [-maxi <max iters>] -knownB0
<bool> [-t <type>] [-i <dwi>] [-o <nout>] [-eo <filename>] [-co
<filename>] [-io <filename>]

no cap, that's the whole vibe right there. notice how -m and -knownB0 are the ones with no brackets? yeah. those are the main characters. everyone else is optional. 👑

🎀 OPTIONS 🎀

• -v <verbose> — how chatty do you want this to be? it's an int, and the default is "0". so by default it's giving silent treatment. 💅
• -m <model> — which model to fit, bestie. (string) and here's the cute part: you can slap an optional "b0+" prefix on it to say yes please also save out the B0 image — and that happens independent of whether the B0 was known or had to be estimated, which is governed by -knownB0. so "b0+" is doing its own thing, okay? she's independent. 🦋
• -ns <# starts> — number of random starting points where the fitting gets initialized. (unsigned int) default: "1". one (1) single start, no more, unless you ask. 🙈
• -ml — do ML fitting instead of least-squares fitting. heads up though: this also requires you to set -sigma. it's a package deal, fr fr. 😭
• -sigma <sigma> — the Gaussian/Rician noise parameter. (double)
• -eps <eps> — convergence epsilon. (double) default: "0.01". that's your "are we there yet" threshold. 🚗
• -mini <min iters> — the minimum required number of iterations for fitting. (unsigned int) default: "3". she needs at least three, she's not a quitter. 💪
• -maxi <max iters> — the maximum allowable number of iterations for fitting. (unsigned int) default: "100". after that we are DONE, no matter what. ⏰
• -knownB0 <bool> — tells the program whether the B=0 non-diffusion-weighted reference image is known. pass "true" if it shows up one or more times amongst the DWIs. pass "false" if it has to be estimated along with all the other model parameters. (bool) this one's not optional, so decide, bestie. 🫶
• -t <type> — output type of the model parameters. default: "float". ✨
• -i <dwi> — all the diffusion-weighted images, all together in one 4D nrrd. default: "-". 🌈
• -o <nout> — the output parameter vector image. (string) default: "-". 🎀
• -eo <filename> — give it a filename and it'll save out the per-sample fitting error. by default, no such error is saved. (string) 📝
• -co <filename> — give it a filename and it'll save out the per-sample convergence fraction. by default, no such error is saved. (string) 📈
• -io <filename> — give it a filename and it'll save out the per-sample number of iterations needed for fitting. by default, no such error is saved. (string) 🔢

lowkey obsessed with how all three of -eo, -co and -io say "no such error is saved" by default. that's just what the page says. i don't make the rules. 💀

👀 SEE ALSO 👀

• tend(1) — go read that one too, bestie, iykyk 🦋

and that's the whole page!! 💜 this little icon of a manual page is dated April 2021, and it came to you from the User Commands corner of the world. mwah. slay. periodt. ✨💅🎀