公开(公告)号：US20130058574A1

公开(公告)日：2013-03-07

申请号：US13574399

申请日：2011-01-21

申请人： Francois Malgouyres ,  Nicolas Lerme ,  Lucas Letocart

发明人： Francois Malgouyres ,  Nicolas Lerme ,  Lucas Letocart

IPC分类号： G06K9/34

CPC分类号： G06K9/00463 ,  G06K9/4638 ,  G06T7/162

摘要： An image segmenting method includes: reading (202) an image, determining (232) a solution to the problem of maximum flow in a graph including, on the one hand, as vertices, a source, a sink and image points, with each point being assigned a capacity, called a through-capacity, assigning (234), on the basis of the determined solution, a label to each of at least some of the points of the image, and recording the image with the assigned labels in a computer memory. In addition, before determining a solution to the problem of maximum flow, the method includes: determining (212) critical points, for each of which, the points of the image located in a predetermined window applied around the critical point verify a predetermined condition on their through-capacities. The points of the graph include the determined critical points and the inter-point arcs link the neighboring critical points to one another.

摘要翻译： 图像分割方法包括：读取（202）图像，确定（232）解决图中最大流量问题的方案，其中一方面包括作为顶点，源，汇点和图像点，每个点 被分配一个称为直通容量的容量，基于所确定的解决方案，将图标的至少一些点中的每一个的标签分配（234），并将具有分配的标签的图像记录在计算机中 记忆。 另外，在确定对最大流量问题的解决方案之前，该方法包括：确定（212）关键点，对于每个关键点，位于围绕临界点应用的预定窗口中的图像的点验证预定条件 他们的能力。 图中的点包括确定的临界点，并且点间弧线将相邻临界点彼此连接。

公开(公告)号：US08798365B2

公开(公告)日：2014-08-05

申请号：US13574399

申请日：2011-01-21

申请人： François Malgouyres ,  Nicolas Lerme ,  Lucas Letocart

发明人： François Malgouyres ,  Nicolas Lerme ,  Lucas Letocart

IPC分类号： G06K9/34 ,  G06K9/00 ,  G06K9/46

CPC分类号： G06K9/00463 ,  G06K9/4638 ,  G06T7/162

摘要： An image segmenting method includes: reading (202) an image, determining (232) a solution to the problem of maximum flow in a graph including, on the one hand, as vertices, a source, a sink and image points, with each point being assigned a capacity, called a through-capacity, assigning (234), on the basis of the determined solution, a label to each of at least some of the points of the image, and recording the image with the assigned labels in a computer memory. In addition, before determining a solution to the problem of maximum flow, the method includes: determining (212) critical points, for each of which, the points of the image located in a predetermined window applied around the critical point verify a predetermined condition on their through-capacities. The points of the graph include the determined critical points and the inter-point arcs link the neighboring critical points to one another.

摘要翻译： 图像分割方法包括：读取（202）图像，确定（232）解决图中最大流量问题的方案，其中一方面包括作为顶点，源，汇点和图像点，每个点 被分配一个称为直通容量的容量，基于所确定的解决方案，将图标的至少一些点中的每一个的标签分配（234），并将具有分配的标签的图像记录在计算机中 记忆。 另外，在确定对最大流量问题的解决方案之前，该方法包括：确定（212）关键点，对于每个关键点，位于围绕临界点应用的预定窗口中的图像的点验证预定条件 他们的能力。 图中的点包括确定的临界点，并且点间弧线将相邻临界点彼此连接。