TEND-EVQ(1)

💅 TEND-EVQ(1) — like, the manual page bestie

okay so this is the page for tend-evq, one of the tend user commands, and honestly? it's a little baddie. let's get into it 🎀

---

✨ NAME ✨

tend-evq — quantize directions of diffusion

---

🌸 SYNOPSIS 🌸

tend evq [-c <evec index>] -a <aniso> [-ns] [-i <nin>] [-o <nout>]

---

🦋 DESCRIPTION 🦋

So. We're quantizing directions of diffusion. Why, you ask? Because VTK literally does not do multi-dimensional colormaps, bestie. Not a thing it does. So to get the classic XYZ<->RGB coloring we all know and love, we have to quantize the directions of diffusion — usually just the principal eigenvector — and then color them. 💅

Here's the tea though: eigenvector directions are poorly defined in regions of low anisotropy. Like, they're just vibing in the wrong direction, no cap. So the length of the vector (pre-quantization, okay, before we touch it) gets modulated by anisotropy. Which means you have to pick some anisotropy metric. That's the whole reason -a exists. periodt.

---

👑 OPTIONS 👑

• -c <evec index> — which eigenvector are we quantizing, bestie? Here's the rundown:
  • "0" gets you the direction of fastest diffusion — that's the eigenvector associated with the largest eigenvalue
  • "1" or "2" gets you the other two eigenvectors — associated with the middle and smallest eigenvalue respectively
  • it's an int, and the default is "0" (main character eigenvector, obviously)
• -a <aniso> — which anisotropy metric do we scale the eigenvector with? 🧚 important detail: all the Westin metrics come in two versions. Currently supported, and this list is the whole list:
  • "cl1", "cl2": Westin's linear
  • "cp1", "cp2": Westin's planar
  • "ca1", "ca2": Westin's linear + planar
  • "cs1", "cs2": Westin's spherical
  • "ct1", "ct2": GK's anisotropy type (cp/ca)
  • "ra": Basser/Pierpaoli relative anisotropy/sqrt(2)
  • "fa": Basser/Pierpaoli fractional anisotropy
  • "vf": volume fraction = 1-(Basser/Pierpaoli volume ratio)
  • "tr": trace
• -ns — don't attenuate the color by anisotropy 🚫🌈. By default, i.e. when you're not passing this flag, regions with low or no anisotropy come out as very dark colors or straight-up black. So if your image is giving "void" and you didn't want that... this is the flag.
• -i <nin> — the input diffusion tensor volume. default: "-"
• -o <nout> — the output image (floating point) (string), default: "-" ✨

---

🫶 SEE ALSO 🫶

tend(1)

---

little reminder bestie: this is all straight from the source, so don't come for me if your anisotropy metric choice makes your volume look like a moody indie film 💀

---

and that's the whole page, fr fr. this lil slice of the user commands manual is dated April 2021, so go forth and quantize those directions. slay 💖