Photographing People in 3D with Photogrammetry

By on February 13, 2013

[phosphor src=”” atlascount=”8″ width=”480″ height=”360″ autoplay=”true” isinteractive=”true” loop=”true”]


When you don’t have an army of vfx artists at your disposal, you are constantly fighting one big enemy: time. Surprisingly, the time budget is your biggest enemy, even more than the money budget.  When working at the high-end vfx level, the time it takes to do any work in post grows exponentially. It’s similar to the amount of effort required to move from shooting stills to shooting videos and how everything becomes a factor of 10 more expensive to do.  Need to edit out a blemish? In the still world, no problem; at most you have 40 selects from a photoshoot. One click with the heal brush in photoshop and you’re done!  Need to do that on a 3 minute video? That’s 4300 frames. My rule of thumb is that it’s an equivalent jump going from photography to film as it is when going from filming to creating purely 3D realistic content. That’s why movies like The Life Of Pi cost 100 million dollars to make even though it was mostly shot in one location on a blue screen sound stage. It takes a lot of man hours to make something CG. But, don’t despair. That’s where our secret weapon #1 comes into play: photogrammetry!

Lee Perry Smith 3D Photocapture

To put it really simply, photogrammetry is a way of creating a 3D model from a photograph.  You take photos from a bunch of different angles, apply a bunch of science and math, and out comes this pretty realistic 3D model. Just think about that for a second. You click a button. And out comes out a full photorealistic 3D model. Imagine how people first felt when they first saw the first cameras. “Wait, I don’t have to have an artist with a lifetime of experience spend an entire day painting me? I can get a picture immediately?” (okay, the first daguerreotypes took forever but you get the point) This is how a lot actor doubles are done in movies.  But, now this is accessible to us everyday people. How? Well for one, our DSLRs can capture 24 megapixel pictures, We can capture really high resolution pictures which is what this technique really needs to capture.

Capturing Jade Bryce

For our Bud Light Girl proof of concept, Cory ( & I had Jade Bryce stop by the studio to be our model Bud Light surfer chick. Now like the first cameras, there are some rough edges that we’re working through.  For example, right now, we need about 45 different angles to create a full 3D scan. This means we have to have either 45 cameras on hand or ask the model to hold really really really still while we walk around her and capture her. We’re R&Ding ways on how to reduce that load from 45 cameras to 16, and we’ll keep you posted on what we find. For Jade, we asked her to hold really still while we shot her in the studio.  The problem with this is that that no one can hold perfectly still and even millimeter movements starts to cause a blur that will start to make the model look like a “video game”.  For example, take a look at her face compared to her body.

Her face looks like it's "game-mapped"

Her face looks like it’s “game-mapped”


The other problem that you would encounter is that the geometry of the model starts have a lot of ‘noise’.

Her body looks much better with our first capture

Her body looks much better with our first capture

Now once we have our 45 different angles, we need to process these images to create the final 3D model. We’ve experimented with several different photogrammetry techniques and software packages, and we settled on Agisoft Photoscan as the best in class. The Agisoft guys are stellar and keep updating the software and big shout out to Lee Perry Smith who helped trailblaze a lot of this stuff. As I said before, we faked 45 different cameras by going around our model taking pictures one by one. This meant that Cory had to spend a lot of time in Maya to cleanup the assets. Not ideal from a time-budget, but it was still reasonable 4-10 hours of cleanup. Remember, the alternative of creating something of that level of detail and quality from scratch would take an experienced artist an entire month or two.  In a later post, we’ll be going over these cleanup techniques in detail. So now that we have our photographic 3D model, how do we get her to show up on top of a coaster? Next post, we’ll talk about how to very easily create an Augmented Reality app with a drag and drop editor (no programming required!).

1 Comment
  1. Reply


    August 22, 2019

    Would be nice if we could use pictures of the model in various poses as well. No need for them to remain perfectly still while you scan around them, take pictures of them from various angles while they are involved in some activity. The software would need to locate and identify limbs, identify skin folds at elbows etc that may not be present in other pictures etc and build a 3D model based on the pictures.

    Aded to a means to move automatic location matches and place some of your own as well as some fixed reference points such as head, hand, elbow, knee and so on, the software could build a 3D VR model based on what it can see from the pictures. If we could then view it in VR as a proper model we could walk around, it would be far better than the photo album we have of our family now.


This site uses Akismet to reduce spam. Learn how your comment data is processed.