Quantitative methods of edge detection

TLDR: A new edge fitting algorithm is introduced that is derived in the discrete domain, this allows a direct optimization of the operator's performance and better performance with real world pictures and less sensitivity to signal-to-noise ratio.

Abstract: : Most local operators used in edge detection can be modelled by one of two methods: edge enhancement/thresholding and edge fitting. This dissertation presents a quantitative design and performance evaluation of these methods. The design techniques are based on statistical detection theory and deterministic pattern recognition classification procedure. The performance evaluation methods developed include: (a) deterministic measurement of the edge gradient amplitude; (b) comparison of the probabilities of correct and false edge detection; and (c) figure of merit computation. The design techniques developed are used to optimally design a variety of small and large mask edge enhancement/ thresholding operators. A performance comparison is given between these edge detectors. A new edge fitting algorithm is introduced. The new algorithm is derived in the discrete domain, this allows a direct optimization of the operator's performance. The advantages of new algorithm are better performance with real world pictures and less sensitivity to signal-to-noise ratio.