.png){kind=small}
MMpose
A motion‑tracking tool for everyone
Improving movement efficiency, reducing injury risk, and supporting rehabilitation all require careful attention to body positions during motion. Traditionally, movement patterns are analysed from photos or videos, but the human eye does not always detect every detail. More information can be extracted from visual material by using photonics‑based technologies, such as MMPose, which was utilised in the TestLab Mehtimäki project.
Download and install our free MMPose installation package (In Finnish only) below to start analysing your own videos and images. The software runs on Windows (version 10 or later). Please review the installation and user instructions included in the package carefully before installation.
If you have ideas or questions about using MMPose, you are welcome to contact us. We value your insights on potential applications of the tool in different contexts and are happy to discuss further:
pasi.vahimaa@uef.fi.
More information about MMPose (in English):
https://mmpose.readthedocs.io/en/latest/overview.html
1
Cost-efficient and powerful
MMpose is a free and versatile tool that detects the key landmarks of the human body from a video or photograph (e.g., elbows and knees). Based on these points, it generates a clear visual stick-figure model overlaid on the person’s image. This enables rapid analysis of movements and postures without the need for markers attached to the body.
2
Easy-to-use, easy-to-move
The use of MMpose requires only a modern computer and a digital camera. This simple equipment setup makes the technology portable and enables analyses to be carried out in a variety of environments and conditions (see, for example, golf). During recording, the person should be shown in full, with no other people in front of or behind them. Lighting conditions should be even, avoiding scenes that are too dim or overly bright. The comparability of images can be improved by using the same devices and maintaining a consistent camera angle for each shot.
3
Open source
MMpose is an open-source program whose code can be freely modified to meet the needs of sports, coaching, or rehabilitation, for example. The model can be extended to include the display of various body angles, which may be useful for technique training in a given sport or in rehabilitation. In addition to the visual model, the program also stores tabular data, enabling more detailed analysis and presentation of the information using, for instance, spreadsheet software or AI-assisted methods.
4
Enables real‑time motion tracking.
MMpose’s source code can also be further developed for use in live situations. To illustrate this, the attached installation package includes a user-friendly program called BodyAngle, which was originally developed as part of the TestLab Mehtimäki project to support the presentation of MMpose to rehabilitation and physiotherapy professionals. In BodyAngle, the user can select the desired body angles with a single button. The program captures an image of the person visible in the webcam, overlays a stick-figure body model on top of the image, and displays the values of the selected angles numerically.