Document Type
Lecture
Publication Date
10-26-2011
Abstract
Fitting geometric curves such as lines, circles, and ellipses to data is a fundamental task in computer vision, image processing, and pattern recognition. These problems are classified as Errors-in-Variables models (EIV), where both coordinates of observations are subject to noise, making them more challenging than classical regression. This talk explores statistical methods to assess EIV parameter estimators, comparing geometric and algebraic fits for circles and ellipses. We present new estimators with superior accuracy and bias reduction.
Relational Format
presentation
Recommended Citation
Al-Sharadqah, Ali, "Statistical Analysis of Curve Fitting in Errors-In-Variables Models" (2011). Analysis Seminar. 12.
https://egrove.olemiss.edu/math_analysis/12
