A saliency estimation model for egocentric motion patterns considering receptive field properties of early visual system and higher order motion detection mechanisms

DOI

Bibliographic Information

Other Title
  • 初期視覚系受容野特性と高次運動選択機構を考慮した自己運動映像に対する顕著性推定モデル

Abstract

The purpose of this study is to construct a mathematical model which predicts saliency regions in high-speed egocentric motion movies. The proposed model was considered the spatial differentiation by the center-surround organization of receptive fields, time response characteristics which generate visual adaptation, and motion feature extraction by higher order motion detection mechanism in the middle temporal area(MT) and the medial superior temporal area(MST). We simulated by using high-speed egocentric-motion movies, filmed by an embedded camera in a driving vehicle, and compared it with human scan-path data. The simulation results indicate that the proposed model detects more salient objects around the vanishing point than the conventional saliency based model. The moving-NSS scores, the index of prediction-accuracy of focal attention area, are significantly higher in our model than those of the conventional saliency estimation models. These results suggest that the proposed model is able to predict the focal attention area similar to the human visual perception in the high-speed self-motion image.

Journal

  • ITE Technical Report

    ITE Technical Report 41.09 (0), 13-16, 2021

    The Institute of Image Information and Television Engineers

Keywords

Details 詳細情報について

  • CRID
    1390851098293998592
  • NII Article ID
    130008043760
  • DOI
    10.11485/itetr.41.09.0_13
  • ISSN
    24241970
    13426893
  • Text Lang
    ja
  • Data Source
    • JaLC
    • CiNii Articles
  • Abstract License Flag
    Disallowed

Report a problem

Back to top