| 摘要 |
The invention relates to a computer-implemented method and system for detecting an object marked by means of stroke input of a user on the basis of pixels in a digital image, which comprises the following method steps: obtaining a digital image, in the memory of a computer; representing the digital image on a display of the computer; receiving stroke inputs of a user via an input interface of the computer; determining the pixels in the digital image which correspond to the stroke input of the user and which are determined to be pertaining to the object to be detected; determining the pixels which in a selectable perimeter region are located around the pixels which correspond to the stroke input of a user, carrying out the object detection by using the pixels which correspond to the stroke input of the user and the pixels in the perimeter region, wherefore a determination is made for each pixel within the perimeter region to determine whether said pixel is located closer to the outer edge of the perimeter region or closer to the pixels that correspond to the stroke input of the user with regard to a selected metric, wherein a pixel within the perimeter region is determined to be associated with the object to be detected, if said pixel with the regard to the selected metric is located closer to the pixels which correspond to the stroke input of the user or alternatively a pixel within the perimeter region is not determined to be associated with the object to be detected if said pixel is located closer to the outer edge with regard to the selected metric, all pixels of the object marked by means of stroke input of the user are detected. |
