University of Twente Student Theses

Login

Markerless stride segmentation for canine pose estimation

Heuver, J. (2022) Markerless stride segmentation for canine pose estimation.

Full text not available from this repository.

Full Text Status:Access to this publication is restricted
Abstract:Pose estimation can deliver valuable insights into the health of animals. The pose can be assessed as is by a professional or further analysed automatically to detect irregular gait patterns for instance. This information can aid in detecting illnesses early on. Recent work by [14] has shown that one of the most robust features to use when using objective gait analysis for lameness detection, is head-tracking combined with footstep detection. The head moves up and down in a sinusoidal pattern, twice per stride, following the steps and loading the extra weight of the head onto the legs. If one of the legs is injured, the animal will avoid loading it, and as such an asymmetry can be seen. Since dogs tend to react badly to attached sensors, and optical motion capture setups tend to be expensive and time consuming to calibrate, markerless motion capture methods such as keypoint estimation could be an invaluable resource by making gait analysis more readily available. Extracting the aforementioned features has not been done yet for dogs so far, to the best of our knowledge, nor from markerless pose data. This is why we pose the following question: can we apply the kinematic feature extraction from keypoint data gathered from dogs; and if so, how well can the feature be used to predict dog lameness?
Item Type:Essay (Bachelor)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Computer Science BSc (56964)
Link to this item:https://purl.utwente.nl/essays/92025
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page