| 摘要 |
The invention relates to a method for segmenting an image of an eye, represented by a set of pixels, comprising a segmenting step providing a segmenting mask comprising a predetermined number P of regions and a step for validating said obtained segmenting mask. The validation of the obtained segmenting mask comprises a step (42) of associating each pixel of the image with a classification identifier, the number N of classification identifiers being strictly greater than P, through an unsupervised classification method, and a step (44) of determining a validation mask comprising P regions, comprising, for at least one region of said segmenting mask, selecting at least one classification identifier associated with a majority of pixels of said region and merging all of the pixels associated with said at least one selected classification identifier to form a corresponding region of the validation mask, and lastly a step (46) of evaluating the segmenting mask using the determined validation mask. The invention also relates to an associated device and a biometric data processing system implementing at least one image of an eye of an associated individual. |
