Reference photos
Importing reference photos as image-plane cameras, aligning them to the head mesh in the Photos window, and grooming through them.

The Photos tab in the Groom Tools creates a camera for each reference photo, aligned to the head mesh, so you can groom while looking through it.
Import a photo
Import Photo... creates a camera aimed at the head with the photo on its image plane. The photo is copied next to the scene, so it moves with the scene.
Look through it
Look Through shows the photo's camera in the active viewport. Click it again to return to the previous camera.
Line it up
Align Photo... opens the Photos window on the photo: the photo on the left, the head mesh on the right, like the Transfer Groom window. You never have to look through the camera to do it.
- Click a feature on the photo (the corner of an eye, the tip of the nose, an ear), then the same feature on the mesh, or the other way round. The Nth point on the photo pairs with the Nth on the mesh, numbered on both sides. Drag a point to adjust it; Ctrl+click deletes it and its partner.
- From four pairs the camera lines up as you place them (the focal length kept). Six or more, spread in depth, solve the focal length too, unless Keep focal length is on.
- If you know the lens, set Focal Length (mm): Keep focal length turns on and the camera lines up again keeping it.
The mesh's outline draws over the photo through the camera, and each pair shows its error in pixels (a line to where its mesh point lands). A pair that disagrees with the rest turns red: fix that one first. Ctrl+Z and Ctrl+Y undo and redo in the window. Below Align Photo..., the tab says how far along the photo is.
You can also line it up by eye: look through it and move the camera.
Morph the scalp to the photos (optional)
When the head mesh is not the person in the photos, Morph to Photos (in the Photos window, under Morph) shapes the scalp so every lined-up photo's pairs land on their photo points. A pair seen in one photo moves across that photo's view; a pair seen in two or more photos gets its depth from them too. The rest of the mesh follows smoothly.
- Symmetry (on by default) morphs both sides alike, each pair mirrored across the scalp's middle.
- Stiffness keeps more of the scalp's own shape: 0 lands every pair exactly, higher moves the scalp more as one piece and the pairs fall short.
- Each press starts from the scalp as it was before the first morph, so pressing it again after moving a pair never stacks morphs. Reset Morph puts the scalp back (undoable).
In Maya the morph is a blendShape on the scalp; in Blender (the Photos tab) a shape key over the Basis. In the app the groom moves onto the morphed scalp by its UVs.
Lock and groom
Lock the camera once it is aligned (in the tab or the Photos window): the landmarks no longer move it, nor does tumbling through it. Set the photo's Opacity, place it In front of the mesh or Behind it, and brush the guides to match the photo while looking through the camera.
Zoom in without moving the camera. Through a locked photo the viewport's navigation zooms and pans the picture in 2D, the photo and the groom together: the wheel and Alt+right drag zoom, Alt+middle drag pans, F shows the whole view again. The camera stays where you lined it up, and the brushes work in the zoomed view. Unlocked, the same gestures move the camera, to line it up by eye. In Blender, a camera view always does this: the wheel zooms, Shift+middle drag pans, Home fits the view.
import groomlab.reference as ref
cam = ref.import_reference("C:/photos/front.jpg", groom=groom)
ref.set_pair(cam, 0, image=(412, 230), mesh=(0.0, 9.5, 3.1)) # ... 6 pairs (4 keeping the focal length)
ref.solve(cam) # the camera from the pairs (the Photos window does it as you click)
ref.morph_to_photos(groom, stiffness=0.0, symmetric=True) # optional; ref.reset_photo_morph(groom)
ref.look_through(cam)
ref.lock_reference(cam)
ref.set_reference_opacity(cam, 0.4)
ref.zoom_view_2d(cam, 3.0, at=(120, 40)) # the view zoomed 3x about a point (pixels from its middle)
ref.reset_view_2d(cam)Guides from photos (experimental)
Experimental
Guides from Photos grows guides that follow the hair in your reference photos. It is a first version: try it, then brush the result.
GroomLab > Guides from Photos (experimental)... opens the Photos window at its Guides stage, or at Align while no photo is lined up yet (Groom Tools > Photos has an Open Guides from Photos... button too; the Align and Guides (experimental) chips at the top left switch stages). Each stage shows only what its step needs:
- Align lines one photo up at a time: the landmarks, the mesh's outline over the photo (Mesh over photo under it), and in the panel the landmark pairs, the camera and the scalp's shape (Morph to Photos). Next: Guides moves on.
- Guides works with every lined-up photo: the tools Guide Lines, Hair and Backdrop above the photo, what it shows below it (Mask, Flow field and the generated Guides, root green to tip red), the groom in the mesh view (Maya), and in the panel this photo's mask and lines, then Generate's settings and the measured look. A photo that is not lined up yet says so in its tab, and Guides from Photos leaves it out.
Each photo keeps its camera with the photo on an image plane, lined up, so renders and Arnold can use it.

Add the photos
Import Photo... adds a photo as a tab along the top (Remove takes it away; Maya's undo brings it back). One to three views work: a front, a side and a back, or a single three-quarter view with symmetry. Each photo has its own landmarks, guide lines, mask and camera. In the photo view the wheel zooms at the cursor, the middle button (or Alt+drag) pans, and F or Fit fits the photo.
Line each photo up with landmarks
At the Align stage, line each photo up as in Line it up above. The mesh can draw over the photo as its outline or its wireframe (Mesh over photo under the photo, with its Opacity).
Draw guide lines (optional)
At the Guides stage, with Guide Lines on, drag along the hair on the photo from root to tip, as one strand runs: the line draws green at its root and red at its tip. Guides near it follow its course and its way, and take its length: a guide growing under a line is as long as the line from there to its tip. Short lines over a beard give short guides; one long line down the back gives long ones. Ctrl+drag erases. Clear Lines forgets them.
Touch up the mask (optional)
The guides grow where the photo's hair mask says there is hair: automatic on a plain backdrop, or an image you pick with Mask (white where the hair is). The automatic mask finds the hair by its colors, learned from the hair around the top of the head, so a face, a neck and clothes are left out even when the scalp is the whole head; a beard the hair's color counts as hair (paint it out with Backdrop to leave it bare). Show > Mask tints it over the photo and Flow field draws the strand directions read from it. Paint with Hair where it missed hair and with Backdrop where it took backdrop for hair (Ctrl swaps them; Brush, beside them, sets the size).
Every landmark, guide line and touch-up is one undo step: Ctrl+Z and Ctrl+Y in the window (or Undo and Redo) walk the current photo's edits.
Generate
Set Symmetry (Auto uses the scalp's own symmetry), Guide count and Max length (Auto: as far as the photos show the hair reaching), uncheck Use this photo on a photo to leave it out, then Generate. The guides go into the fromReference group and are selected, and they draw over each photo (Show > Guides, root green to tip red; in Maya on the scalp in the mesh view too): where they stray from the hair, draw guide lines there and Generate again. Generate again replaces them; one Maya undo step each time. While it works, the window shows a wait cursor, greys out its buttons and says what it is doing (Measured Look's Apply too).
Measured Look shows what the photos measure against what your graph has: the hair color (and lighter or darker tips), clumps (how far apart and how tight), curls, flyaways, frizz, and where the hair thins. Uncheck what you don't want and click Apply: the values are set in one undo step, and a Clump, Curl, frizz Noise or Flyaway node the graph lacks is added where it usually goes. Thinning becomes a vertex map (refDensity) on Scatter's density.
These are rough measures from the picture, a starting point to tune by eye. Some things can't be read from a photo yet: each strand's own color variation (it hides in the shading), frizz in curly hair, clumps in short fur, and flyaways against a backdrop of the hair's own shade.
The standalone app has the same window: Align Photo... in its Photos tab, and Tools > Guides from Photos (Experimental)... (photos, landmarks, the camera, guide lines, mask touch-ups and Generate, saved with the groom; the measured look is Maya's for now).
What works so far: long hair and longer fur, one to three views, symmetry that covers the side a photo doesn't show, and guide lines that settle which way hair runs and how long it is.
What doesn't yet: tight curls in three or more views, very short fur (too fine to read in the photo), and photos lit by colored light (the fitted color takes the light's tint).
import groomlab
from groomlab import from_reference as fr
from groomlab.ui import reference_window
w = reference_window.show(groom, stage="guides") # the Photos window; scripts drive it too:
cam = w.import_photo("C:/photos/front.jpg")
w.add_landmark("photo", (412, 230)); w.add_landmark("mesh", (0.0, 9.5, 3.1)) # ... 4 pairs or more
w.add_flow_stroke([(820, 300), (790, 620)]) # photo pixels, root first
fr.set_reference_hints(groom, cam, [[(820, 300), (790, 620)]]) # or straight onto the groom
fr.set_mask_strokes(groom, cam, [{"points": [(100, 80), (180, 90)], "radius": 12, "include": False}])
result = fr.generate(groom, guide_count=400, symmetry="auto")
fr.measured_look_preview(groom) # [{"type", "node", "param", "current", "value"}, ...]
fr.apply_measured_look(groom) # or types=["HairColor", "Clump"]Scripts can also pass a painted overlay: a PNG at the photo's size with the strokes opaque on a
transparent layer, dark at the root end and light at the tip ({"camera": cam, "hint_image": "flow.png"} in groomlab.guides_from_reference).
Motion & simulation
Procedural motion with the Wind and Jiggle nodes, guide simulation with the Simulate node, follow-the-scalp evaluation, nHair in Maya, guide caches from other packages, and shot correctives.
Importing XGen and Yeti grooms
Turning Maya XGen descriptions (Interactive Groom Splines and XGen Core) and Yeti graphs into GroomLab grooms, what maps to what, and the limits.