Segmentation
Basic Segmentation
The Surface Module has two algorithms to segment medical images.
The first is “Basic” which contains three methods:
Threshold – binarize an input image by thresholding. In this method, lower and upper threshold values are set by the user to suit him. There is a special option named “only connected”. If selected, a point can be put in the image series. Then only values in the range of threshold and connected to the point will be taken into creating a mask.
Automatic Histogram Analysis – automatic calculation of threshold by dividing the image into the specified number of upper and lower threshold values. The number of created masks is the same as the specified number, but the user can uncheck the ones he does not want to create.
Point & Grow– interactive mask area expansion based on neighbourhood analysis. This method does not use a threshold, but a starting point. The user chooses a location to seed a point and grows it until the entire structure of interest is filled. One arrow makes the growing up by one iteration, while a double arrow makes it by five.
The second algorithm is called “Smart” . This method starts with cropping a volume (this can be skipped, but it is not recommended) to reduce the workspace to significantly speed up the entire process. If a mistake is made with cropping, it can be easily undone by clicking the revert arrow button. The next step is defining a foreground (what is the area of interest) and a background (what is not the area of interest). During the whole process, at any time the user can check how the mask looks by means of Preview 2D, Preview 3D and Show 3D Surface Preview.
Would you like to start working
with Customy Vision?