Self-Adaptive Thresholding And Its Application To Extracting Nuclei In Glomeruli Of Human Kidneys
X. Zhang, K. Taniguchi, Yuya Nakano · International Journal of Modelling and Simulation · 1997
The authors propose a feature-feedback variable thresholding method in which feature information of the thresholded result is used to tune the thresholding in a proper state where an optimal thresholded result can be obtained, and thus one that is close-loop variable thresholding different from the usual variable thresholding. An original image is convolved with a two-dimensional Gaussian function, and the convolved (blurred) image is used as a surface threshold at which the original image is thresholded. For the thresholded binary image, the selected feature value of extracted objects is calculated, and the value is fed bade to the set value known in advance. The space constant σ of the Gaussian is adjustedaccording to the error between the two values recursively until a proper result is given. The proposed method is applied to extracting nuclei in the glomeruli of human kidneys, where feature selection and the relation between the selected feature and the space constant σ of the Gaussian are discussed. Experimental results are presented to demonstrate the feasibility of the proposed method.