Optical motion analysis

3D image with two skeletonized people
© Fraunhofer IMS
3D image with two skeletonized people

Motion analysis generates valuable information about the activity behavior and execution of movements of humans as well as animals. This also includes algorithms that can be used to detect falls. Up to now, acceleration sensors attached to the joints have mainly been used for this purpose. These are uncomfortable to wear over the long term and can cause skin irritation, especially in older people after prolonged wear. Another solution is marker-based motion analysis using optical sensors. For this, corresponding markers are attached to the clothing and can be detected and tracked via a camera. While this method is also not completely contactless and is unusable for everyday use, artificial intelligence algorithms enable complete marker less optical motion analysis.

Deep neural networks can generate skeletal models by recognizing, associating, and tracking individuals and their joints and body parts. Fraunhofer IMS focuses on the following research questions, among others.

Optical gait analysis:

Using this movement data, gait parameters can be extracted to make an assessment of the risk of falls of elderly persons. Since small irregularities and deviations in the gait pattern are important here, the positions of the joints must be determined as accurately as possible. How accurately the measurements can be realized with a conventional RGB camera, a 3D depth camera and a combination of both is a current research topic at Fraunhofer IMS.

Motion analysis for activity detection:

In addition, skeletal models combined with other algorithms can provide information about a person's pose and posture. This is particularly useful for research on the automatic long-term assessment of health parameters, such as orthostatic intolerance. Here, simultaneous measurement of blood pressure and heart rate can be used to evaluate the autoregulatory power of the cardiovascular system during different activities.

Motion analysis for fall detection:

A detection and classification of different poses is also relevant for other application areas. Another related research topic in the field of home care and geriatrics is fall detection. Basing on pose and activity recognition it can be evaluated whether a transition between poses was normal and intentional or whether it was a fall. In addition, Fraunhofer IMS is working on detecting falls before they happen. By analyzing fall risks, drastic health damage and subsequent costs can thus be prevented.

Our technologies - Innovations for your products

Person Recognition

Special feature extraction is used to enable the recognition of persons on small embedded systems.

Quantification of Health Parameters

Computer vision enables non-contact measurement of health parameters

Contactless measurement

Reduced risk of infection and relief for medical staff thanks to innovative contactless measurement technology

Our technology areas - Our technologies for your development

Communication and Networking

Communication interfaces allow data exchange with other devices and connection to networks

User Interfaces

User interfaces as interface between device and user allow the configuration and operation of a product

Machine Learning for Embedded Systems

Artificial intelligence on resource-limited systems can be used to extract higher quality information from raw sensor data

Computer Vision

Computer Vision methods extract the maximum amount of information from image data

 

Embedded Software and Artificial Intelligence (Home)

Here you can get back to the overview page of the core competence Embedded Software and Artificial Intelligence (ESA).