Generation of Multiple Background Model by Estimated Camera Motion Using Edge Segments

Authors:
Taeho Kim;Kang-Hyun Jo
Affiliations:
Graduate School of Electrical Engineering, University of Ulsan, Ulsan, Korea 680 - 749;Graduate School of Electrical Engineering, University of Ulsan, Ulsan, Korea 680 - 749
Venue:
ICIC '08 Proceedings of the 4th international conference on Intelligent Computing: Advanced Intelligent Computing Theories and Applications - with Aspects of Theoretical and Methodological Issues
Year:
2008

Citing 7
Cited 0

Pfinder: Real-Time Tracking of the Human Body

IEEE Transactions on Pattern Analysis and Machine Intelligence
Image Sequence Analysis via Partial Differential Equations

Journal of Mathematical Imaging and Vision
Detected motion classification with a double-background and a neighborhood-based difference

Pattern Recognition Letters
Multiple Motion Scene Reconstruction with Uncalibrated Cameras

IEEE Transactions on Pattern Analysis and Machine Intelligence
Distinctive Image Features from Scale-Invariant Keypoints

International Journal of Computer Vision
Detecting Motion Regions in the Presence of a Strong Parallax from a Moving Camera by Multiview Geometric Constraints

IEEE Transactions on Pattern Analysis and Machine Intelligence
Statistical modeling of complex backgrounds for foreground object detection

IEEE Transactions on Image Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

We investigate new approach for segmentation of moving objects and generation of MBM (Multiple Background Model) from an image sequence by a mobile robot. For generating MBM from unstable camera, we have to know the camera motion. When we correlate two consecutive images to calculate the similarity, we use edge segments to reduce computational cost. Because the regions, neighbors of edge segments, have distinctive spatial features while some regions like blue sky, empty road, etc. have ambiguity. Based on the similarity result, we obtain best matched regions, their centroids and displacement vector between two centroids. The highest density of displacement vector histogram, named motion vector, indicates camera motion between consecutive frames. We generate MBM based on motion vector and MBM algorithm classifies each matched pixel to several clusters. The experimental results shows that proposed algorithm successfully detect moving objects with MBM when camera has 2-D translation.