Estimation of FOE without optical flow for vehicle lateral position detection

Tomoaki Teshima, Hideo Saito, Shinji Ozawa, Keiichi Yamamoto, Tohru Ihara

Research output: Chapter in Book/Report/Conference proceedingConference contribution

10 Citations (Scopus)

Abstract

In this paper, we propose a new method to estimate the FOE (Focus of Expansion) from moving camera image sequence for vehicle motion understanding, which does not rely on optical flow vectors. Motion understanding of the camera is very important in many applications. Most of the conventional methods for detection and estimation of the position of FOE rely on estimation of optical flow vector in the input image sequence. However, errors in estimating optical flow vector result in inaccurate estimation of FOE. In our method, we synthesize the input image and create an expected image of the next frame by using optical flow vectors. Optical flow vectors are obtained from the hypothesis of position of FOE and global constraint of optical flow. We evaluate the position of FOE by SAD (Sum of Absolute Differences) value between the warped and the real image. The proposed method estimates the position of FOE by minimizing SAD. For demonstrating the efficacy of the proposed method, we apply the proposed FOE estimation method for detecting lateral position of a vehicle. The experimental results show advanced performance comparing with FOE estimation based on a commercial optical flow vector field estimating software.

Original languageEnglish
Title of host publicationProceedings of the 9th IAPR Conference on Machine Vision Applications, MVA 2005
Pages406-409
Number of pages4
Publication statusPublished - 2005 Dec 1
Event9th IAPR Conference on Machine Vision Applications, MVA 2005 - Tsukuba Science City, Japan
Duration: 2005 May 162005 May 18

Publication series

NameProceedings of the 9th IAPR Conference on Machine Vision Applications, MVA 2005

Other

Other9th IAPR Conference on Machine Vision Applications, MVA 2005
CountryJapan
CityTsukuba Science City
Period05/5/1605/5/18

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition

Fingerprint Dive into the research topics of 'Estimation of FOE without optical flow for vehicle lateral position detection'. Together they form a unique fingerprint.

  • Cite this

    Teshima, T., Saito, H., Ozawa, S., Yamamoto, K., & Ihara, T. (2005). Estimation of FOE without optical flow for vehicle lateral position detection. In Proceedings of the 9th IAPR Conference on Machine Vision Applications, MVA 2005 (pp. 406-409). (Proceedings of the 9th IAPR Conference on Machine Vision Applications, MVA 2005).