Multibody motion segmentation for an arbitrary number of independent motions

Yutaro Sako,Yasuyuki Sugaya

doi:10.1186/s41074-016-0002-3

Abstract

Abstract We propose a new method for segmenting feature point trajectories tracked through a video sequence without assuming a number of independent motions. Our method realizes motion segmentation of feature point trajectories by hierarchically separating the trajectories into two affine spaces in a situation that we do not know the number of independently moving objects. We judge that input trajectories should be separated by comparing the likelihoods computed from those trajectories before/after separation. We also consider integration of the resulting separated trajectories for avoiding too much segmentations. By using real video images, we confirmed the efficiency of our proposed method.

Highlights

Separating independently moving objects in a video sequence is one of the important tasks in computer vision applications
Vidal et al [7] proposed a segmentation algorithm based on generalized principal component analysis (GPCA) [6]
Kanatani and Matsunaga [2] proposed a method for estimating the number of independently moving objects based on the rank estimation of the affine space using the geometric minimum description length (MDL)

Summary

Introduction

Separating independently moving objects in a video sequence is one of the important tasks in computer vision applications. Kanatani and Matsunaga [2] proposed a method for estimating the number of independently moving objects based on the rank estimation of the affine space using the geometric minimum description length (MDL). Estimating the number of independently moving objects based on the rank of the affine space is very difficult for real image sequences.

Results

Conclusion