Define Region of Interest

After adding a tool to process the images, they will be blurred and indicate that they need to be segmented by the message, "No ROI (Region of Interest) defined." In order to make these images available to the current tool, you have to segment them by clicking the tool's process button - the Scissor icon.

The default segmentation provides the tool with the complete image. Alternatively, you can manually adjust the ROI (right-click within the image and select Edit ROI) to configure the extracted view in terms of its size, position and orientation. In many cases, you need to adjust the Region of Interest (ROI) to define the view that the tool will operate within. You can adjust the size and placement of the ROI using the graphic handles in the image. This allows you to re-size the ROI, and to move it within the image. When configuring your Region of Interest (ROI), attempt to always create a ROI that is as small as possible.

When the first tool is added, the default Region of Interest toolbar will appear at the top of the image (you can also open it by selecting Edit ROI from the right-click menu).

Expert Mode version:

Setting	Description
Percent/Pixel	Specifies if the ROI will be constructed as a percentage of the width and height of the image, or in the pixel dimensions. Note: If the preceding tool is a Blue Locate tool, the percentage will be based on the width and height of the feature or node model that is defining the view.
Offset	Specifies an offset of the upper left corner of the ROI from the origin. The setting is either a percentage or in pixels, depending on the Percent/Pixel setting.
ROI Size	Specifies the size of the ROI, either as a percentage or in pixels, depending on the Percent/Pixel setting.
Angle	Specifies an angular orientation, from the origin, of the ROI.
Grid	Specifies that the image will be divided into smaller views, based on the grid settings. This is useful when you are working on large images that may contain a large number of features and/or regions, and dividing the image into multiple, smaller views will make it easier to visualize and label. Allowing the Deep Learning application to divide the image versus altering the image in another photo editing program also prevents artificial artifacts from being introduced, because within Deep Learning, the image just continues. Also, when processing images during runtime, the Grid setting can be reset to the default (1 x 1), and the entire image will be processed at once. For example, if you had a large image, which contained 10,000 small items that will be labeled and identified by a Blue Locate tool. By dividing the image into a 10 x 10 grid, you would only need to identify and label 100 features per image, versus 10,000 in one image. Note: This option is only available in Expert Mode.
External	Specifies that the ROI will be oriented by an external source, such as a pose from a VisionPro CogPMAlignTool. Note: This option is only available in Expert Mode.

Note: For chained tools, the ROI is dependent on the preceding tool, and will be dynamic, in other words, the view will depend on the preceding tool's result. It means that if the tools are chained, the preceding tool will determine the available ROI toolbar options. For more information, refer to the .