A Bio-Inspired Computational Model for the 3D Motion Estimation

被引：0

作者：

Zhang, Di ^{[1
]}

Ping, Lu ^{[2
]}

Qi, Wu ^{[1
]}

Ye, Long ^{[1
]}

机构：

[1] Commun Univ China, Beijing, Peoples R China

[2] Beijing Univ Posts & Telecommun, Beijing, Peoples R China

来源：

2024 IEEE CONFERENCE ON VIRTUAL REALITY AND 3D USER INTERFACES ABSTRACTS AND WORKSHOPS, VRW 2024 | 2024年

基金：

国家重点研发计划;

关键词：

Stereoscopic motion perception; Stereo Motion Perception Model (SMPM); Directional asymmetry; Human visual system characteristics;

D O I：

10.1109/VRW62533.2024.00249

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Stereoscopic motion perception is pivotal for human-world interaction and has sparked advancements in artificial vision technologies. We introduce the Stereo Motion Perception Model (SMPM), a bio-inspired computational model designed to emulate the process of stereoscopic motion perception inherent in the human visual system. The SMPM extracts temporal, spatial, and shape features from stereoscopic video stimuli, enabling accurate motion perception. The performance of SMPM is tested and compared with human perception. The results indicates that the SMPM effectively estimates target motion on both simplistic and complex scene, high consistency is shown with human stereo motion perception. Future implications of the bio-inspired model is discussed.

引用

页码：905 / 906

页数：2

共 10 条

[1] Toward the next-generation VR/AR optics: a review of holographic near-eye displays from a human-centric perspective [J].