A new method proposed to explore the feline's paw bones of contributing most to landing pattern recognition when landed under different constraints

Datao Xu; Huiyu Zhou; Qiaolin Zhang; Julien S Baker; Ukadike C Ugbolue; Zsolt Radak; Xin Ma; Fekete Gusztav; Meizi Wang; Yaodong Gu

doi:10.3389/fvets.2022.1011357

A new method proposed to explore the feline's paw bones of contributing most to landing pattern recognition when landed under different constraints

Front Vet Sci. 2022 Oct 10:9:1011357. doi: 10.3389/fvets.2022.1011357. eCollection 2022.

Authors

Datao Xu^{1

2

3}, Huiyu Zhou^{1

4}, Qiaolin Zhang¹, Julien S Baker⁵, Ukadike C Ugbolue⁴, Zsolt Radak⁶, Xin Ma⁷, Fekete Gusztav^{2

3}, Meizi Wang^{1

8}, Yaodong Gu¹

Affiliations

¹ Faculty of Sports Science, Ningbo University, Ningbo, China.
² Savaria Institute of Technology, Eötvös Loránd University, Szombathely, Hungary.
³ Faculty of Engineering, University of Pannonia, Veszprem, Hungary.
⁴ School of Health and Life Sciences, University of the West of Scotland, Scotland, United Kingdom.
⁵ Department of Sport and Physical Education, Hong Kong Baptist University, Kowloon, Hong Kong SAR, China.
⁶ Research Institute of Sport Science, University of Physical Education, Budapest, Hungary.
⁷ Department of Orthopedics, Huashan Hospital, Fudan University, Shanghai, China.
⁸ Faculty of Health and Safety, Óbuda University, Budapest, Hungary.

Abstract

Felines are generally acknowledged to have natural athletic ability, especially in jumping and landing. The adage "felines have nine lives" seems applicable when we consider its ability to land safely from heights. Traditional post-processing of finite element analysis (FEA) is usually based on stress distribution trend and maximum stress values, which is often related to the smoothness and morphological characteristics of the finite element model and cannot be used to comprehensively and deeply explore the mechanical mechanism of the bone. Machine learning methods that focus on feature pattern variable analysis have been gradually applied in the field of biomechanics. Therefore, this study investigated the cat forelimb biomechanical characteristics when landing from different heights using FEA and feature engineering techniques for post-processing of FEA. The results suggested that the stress distribution feature of the second, fourth metacarpal, the second, third proximal phalanx are the features that contribute most to landing pattern recognition when cats landed under different constraints. With increments in landing altitude, the variations in landing pattern differences may be a response of the cat's forelimb by adjusting the musculoskeletal structure to reduce the risk of injury with a more optimal landing strategy. The combination of feature engineering techniques can effectively identify the bone's features that contribute most to pattern recognition under different constraints, which is conducive to the grasp of the optimal feature that can reveal intrinsic properties in the field of biomechanics.

Keywords: animal biomechanics; cat paws; feature engineering techniques; feline landing; metaheuristic optimization algorithms; post-processing of finite element analysis.