KEYPOINT IDENTIFICATION

A method for identifying keypoints in a digital image comprising a set of pixels is proposed. Each pixel has associated thereto a respective value of an image representative parameter. Said method comprises approximating a filtered image. Said filtered image depends on a filtering parameter and comprises for each pixel of the image a filtering function that depends on the filtering parameter to calculate a filtered value of the value of the representative parameter of the pixel. Said approximating comprises: a) generating a set of base filtered images; each base filtered image is the image filtered with a respective value of the filtering parameter; b) for each pixel of at least a subset of said set of pixels, approximating the filtering function by means of a respective approximation function based on the base filtered images; said approximation function is a function of the filtering parameter within a predefined range of the filtering parameter; the method further comprises, for each pixel of said subset, identifying such pixel as a candidate keypoint if the approximation function has a local extreme which is also a global extreme with respect to the filtering parameter in a respective sub-range internal to said predefined range. For each pixel identified as a candidate keypoint, the method further comprises: c) comparing the value assumed by the approximation function at the value of the filtering parameter corresponding to the global extreme of the pixel with the values assumed by the approximation functions of the adjacent pixels in the image at the values of the filtering parameters of the respective global extremes of such adjacent pixels, and d) selecting such pixel based on this comparison. La présente invention se rapporte à un procédé qui permet d'identifier des points clés dans une image numérique comportant un ensemble de pixels. Une valeur respective d'un paramètre représentatif d'image est associée à chaque pixel. Ledit procédé consiste à se rapprocher d'une image filtrée. Cette image filtrée dépend d'un paramètre de filtrage et comprend, pour chaque pixel de l'image, une fonction de filtrage qui dépend du paramètre de filtrage afin de calculer une valeur filtrée de la valeur du paramètre représentatif du pixel. Ledit rapprochement consiste : a) à générer un ensemble d'images filtrées de base, chaque image filtrée de base correspondant à l'image filtrée à l'aide d'une valeur respective du paramètre de filtrage; b) pour chaque pixel d'au moi