The official manual can be found here[akiya-research-institute.github.io].

It contains some technological jargon which should be familiar to other scholars in the field, but it does a good job at explaining the steps. What it does not do is tell you how to operate the program, which this guide should hopefully walk you through.

The minimum amount of space you'll need to get a good capture is about 2.5 to 3 meters.
You can have two cameras in a 2.5 x 2.5 area like in the diagram below:



However, I recommend a larger area (3x3 or 4x4 meters) so that you can actually move around, and side to side.

This Rokoko diagram puts it aptly how much room you need ideally[vision.rokoko.com]


You will need at least TWO cameras, that can capture at least SD (560p) and above. Ensure they're connected to the computer.

These cameras should be positioned around chest height (ie. around the height of a webcam on top of a monitor or on a tripod) and looking straight ahead so that it has the whole area in front of the camera in frame.

First, generate an ChArUco board[docs.opencv.org], or simply use the image below:


(Direct)[docs.opencv.org]

Save a copy of the image and display it on your screen as large as you can, or print it onto a piece of paper and paste it onto a wall. It should be around A3 size.

If you do not have the application open, you should open it now.

First, click on "Add Camera", then next to "Camera:", select the arrow.
In the dropdown, pick your first camera.

Ensure the app is seeing your camera in the preview, and pick up your camera. Adjust resolution where needed.

Next, select "Start" in Intrinsic calibration and make sure you have the entire ChArUco board in view. Shake it and move it around a bit until you hear a "ding". This means it is calibrated properly.


Generate some ArUco markers or use these ones from the official guide[github.com].

You'll need to print them identally sized. Depending on the distance of the cameras, you'll need to print them on bigger paper. For the case of 2.5 to 3 meters, A4 is good enough.

Print out at least THREE different markers.

Before doing anything, measure the size of your boxes (on A4 and standard settings, this is around 17-19cm; 0.17-0.19m).

Go to Settings > Calibration, and set "Extrinsic Calibration Method" to "ArUco cluster" and then set "ArUco marker [m]" to the correct box size.

Place your markers on the floor close to each other where you plan to stand; this should be in view of all cameras without having to move them around; about 2.5 meters away from them.
DO NOT move them after placing down as when calibrating the other cameras they must be in the same position.

Click on "Scan Markers", which should hopefully light up now. Give it a few seconds, then click "Start". It should do some calculations and ding again.


Aand repeat for your other cameras. Ensure that you don't move the markers for the second stage (extrinsic matrices) otherwise you'll throw calibration off.

To actually capture, click "Start Capture".

You can hide/collapse the camera controls by clicking the arrow next to "Camera:" but don't close your cameras as it'll unload them.

You should now see at least two cameras appear in the scene.

To move around, hold right click and WASD like in freecam.

Finally move (physically) into full view of the cameras (around where you placed your cluster boards) and start moving. The character should update as soon as the cameras see you.