The results of the investigation of the boaventura and gonzaga integrated performance evaluation method of edge detection based on the two-dimensional renewal stream

The paper presents the results of the investigation of the I. Boaventura und A. Gonzaga integrated performance evaluation method of edge detection [1-2], obtained using the bundled software of stochastic simulation ≪CS sF≫ [3]. The methods and approaches of stochastic simulation were used in the experiments and the reference images were approximated with the two-dimensional renewal stream [4-7]. The performance of the outline drawing detection was evaluated by Boaventura und Gonzaga method for three algorithms of the edge detection (≪ Canny≫, ≪Marr-Hildreth≫ and ≪ISEF≫) under different levels of peak signal-to-noise ratio. The results of the investigation are presented as dependences of the estimate probability of the correct edge detection, the type 1 and 2 errors, on S/N ratio. The performance analysis of the above three algorithms for the images, produced on the basis of morphology type ≪A≫ and ≪F≫, is done based on the performed evaluation.