Research on the Parameter Estimating Algorithms of Image Edge Detection

Abstract:

During detecting the edge of the images, the text partly use great likelihood estimation and least square method estimation algorithm to estimate, we found the result of two estimate algorithms used in the same model are different through experimental analysis. Aiming at above mentioned problems, this text divides the commonly used model in pattern process into the linear model and non-linear model, among the non-linear model, it divides into multinomial model, gauss model, shouldered index model and power counting model, and this text use great likelihood estimate algorithm and least square method estimation algorithm to estimate these models separately, and draw their scope of the application through the experiment, also provide the convenience for the future choice.

• Periodical: Advanced Materials Research (Volumes 562-564)
• Pages: 1279-1285
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.562-564.1279
• Online since: August 2012
• Authors: Ya Ceng Shang, Jing Chen, Jun Wei Tian
• Keywords: Edge Detection, Image Processing, Least Square Estimation, Maximum Likelihood-Estimation (MLE), Sub-Pixel

References

[1] YE Zong-yu. Study on the Parameter Estimation Method for the Nonlinear Regression Model,. Statistics & Information Forum. 2010, vol. 1, pp.41-45.

[2] YE Ming, Gilbert Saporta, WANG Hui-wen. Nonlinear Regression Automatic Modeling Process,. Systems Engineering. 2009, vol7, pp.81-84.

[3] XIE Lan, GAO Dong-hong. The Application and Comparison of Different Nonlinear Fit Methods,. 2009, vol10, pp.117-121.

[4] OUYANG Guang. Modified Maximum Likelihood Estimation of the Shape Parameter for Gamma Distribution,. Journal of Xiangnan University. 2009, vol2, pp.21-23.

[5] DAI Jia-jia, YANG Ai-jun, YANG Zhen-hai. An advanced algorithm for equations of maximum likelihood estimation". University, s applied mathematics journal. 2009, vol3, pp.275-280.