Audio and Video

New motion capture model creates virtual dogs without expensive equipment

16 June 2020

Researchers from the University of Bath created a new way to digitize dogs without expensive motion-capture suits or multiple cameras. This technology would help vets virtually diagnose and monitor lameness in dog patients and digital artists add realistic motion to the dogs in their movies and video games.

When using motion capture technology, actors wear suits dotted with white markers that are tracked in a 3D space by multiple cameras from different angles. The movement data is transferred to a computer and turned into a digital rendition. Similar technology is used by biomechanics experts to track the movement of athletes during training and monitor an injured patient's rehab. Using today’s motion capture technology with animals requires expensive equipment and dozens of markers attached to the animals.

Sinead Kearney adjusts the cameras to collect the motion capture data of a lurcher Source: University of BathSinead Kearney adjusts the cameras to collect the motion capture data of a lurcher Source: University of Bath

To develop the new motion capture technology for animals, the team digitized 14 dog breeds. The dogs were residents of Bath’s Cats’ and Dogs’ Home (BCDH). The dogs wore special motion capture suits with markers and filmed under the supervision of BCDH handlers while they performed a range of motions. Researchers used this data to create a computer model that can predict and replicate the poses of dogs when filmed without the motion capture suits.

The new model allows 3D digital information for new dogs, like shape and movement, to be captured without markers or equipment, using only an RGBD camera. RGBD cameras can record RBG color in pixels in an image and the distance the animal is from the camera for each pixel. This is more affordable than traditional motion capture systems that require multiple cameras.

The team has started testing the method on other four-legged animals with promising results. They aim to extend their animal dataset for more accurate results.

This research was presented at the Computer Vision and Pattern Recognition Conference. Access the research paper here.



Powered by CR4, the Engineering Community

Discussion – 0 comments

By posting a comment you confirm that you have read and accept our Posting Rules and Terms of Use.
Engineering Newsletter Signup
Get the Engineering360
Stay up to date on:
Features the top stories, latest news, charts, insights and more on the end-to-end electronics value chain.
Advertisement
Weekly Newsletter
Get news, research, and analysis
on the Electronics industry in your
inbox every week - for FREE
Sign up for our FREE eNewsletter
Advertisement