A team of researchers from the Centre for Motion, Recreational Research and Applied Analysis at the University of Bath has just demonstrated a computer software model they created that features a single camera and no fancy motion capture kit to digitize dogs. Thanks to this technology, future gamers are expected to easily bring their pets into the digital world and have a positive impact on the film and television industry.
Sinead Kearney adjusts the camera to capture pet movement data (from University of Bath)
It is reported that the research team of 14 different breeds of pet dog operation digital collection. In the early stages of the study, the team put the pet on a special motion-capture suit and then took action images of their actions, trots, or jumps.
They then used the collected data to create a new set of computer models that predicted and mimicked the posture of various breeds of dogs. And 3D data can be used to shoot other pet dogs without the need for special action-gathering costumes.
The new computer model also eliminates the need for expensive digital acquisition and processing equipment, which can be achieved with a single RGBD camera.
Traditional digital cameras record through RGB pixels, and a single RGBD can also record the distance from the captured object to the camera.
Predicting Canine Pose from RGBD Sensors (via)
For the first time, researcher sin?ad Kearney said in a statement that they were using a single RGBD image capture system to track the movements of pet dogs, and that the overall cost was much more affordable than traditional motion capture systems based on multiple cameras.
Looking ahead, the technology is expected to be adopted by budget-sized filmmakers or game developers to introduce more realistic digital animal characters to their projects.
Dog motion capture (via)
Pet owners can also display their pets in 3D on their computers, or help veterinarians monitor the recovery of injuries or illnesses.
Finally, the team presented their latest findings at the Annual Conference on Computer Vision and Pattern Recognition, which was held online on June 14.