Combined optic flow optimization method using Pulse-Couple Neural Network

Authors:
Yanpeng Cao;Peter Cook;Alasdair Renfrew
Affiliations:
University of Manchester, UK;University of Manchester, UK;University of Manchester, UK
Venue:
CGIM '07 Proceedings of the Ninth IASTED International Conference on Computer Graphics and Imaging
Year:
2007

Citing 6
Cited 0

Digital image processing (2nd ed.)

Digital image processing (2nd ed.)
Neural mechanisms of scene segmentation: recordings from the visual cortex suggest basic circuits for linking field models

IEEE Transactions on Neural Networks
PCNN models and applications

IEEE Transactions on Neural Networks
Perfect image segmentation using pulse coupled neural networks

IEEE Transactions on Neural Networks
Motion segmentation based on motion/brightness integration and oscillatory correlation

IEEE Transactions on Neural Networks
Temporally sequenced intelligent block-matching and motion-segmentation using locally coupled networks

IEEE Transactions on Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

Noise disturbance and "Blank Wall" problem are two main challenges faced in optic flow estimation based on Block Matching. In this paper we propose a novel method to optimize optic flow by combining BMP Re-calculation and color image segmentation. Firstly standard Block Matching technique is applied to two consecutive images to produce optic flow estimation of scene; after introducing novel concept -- Block Matching Probability (BMP), 3D Pulse-Coupled Neural Network (PCNN) is applied to re-calculate BMP value of displacements of pixels on boundary; then a novel PCNN is proposed to partition image based on color information; in the end optimized BMP values and image segmentation are combined to produce more accurate optic flow estimations. Experimental results show combined method effectively improves accuracy of optic flow estimation in both synthetic and real images.