Overview Region of Interest (ROI) Node
The Region of Interest (ROI) node allows to filter a portion of an image that you want to perform some other operation on.
ROI can be set for each camera stream and is therefore device specific. While the ROI node needs to be added to your flow, it will be configured in the
Local Configuration settings, not inside the Viso Builder.
To access the local configuration settings and setup your ROIs, follow the guide
here.
Name: You can set a name for your ROI node. This is especially important if you are using multiple camera input streams and multiple ROI nodes in your flow.
ROI Type: The ROI node allows you to select from three types of ROIs, depending on your use case:
- Rectangle: The rectangle ROI type is used for counting use cases. It allows you to define the "counting area" in the image (blue rectangle) and position the crossing line (red). Objects crossing this line will be counted (IN/OUT). If you select "Crop ROI", then the image will be cropped to the ROI size to increase inference performance.
Additionally, the rectangle type allows you to rotate the incoming frames. You will need that for example in the case you are using a fisheye camera and want to count people at the entrance of a room. For optimal performance with the given algorithm, you might want to have the entrance on top of the image and count people crossing the line horizontally. - Polygon: The polygon ROI type allows you to set your region of interest as a polygon. This might be used if you, for example, want to detect objects only in a specific area of a specific shape (e.g. intrusion detection). Once you have set your polygon, you will be prompted with a popup to give your ROI a name. This allows you to have multiple polygon ROI on the same input stream.
Additionally, you will see the option to "exclude an ROI". This option will be helpful if you would like to set certain areas in which no objects should be detected. - Section: The section ROI type allows you to define multiple sections in a video frame. You can flexibly adjust the number of sections, grid, area names and colors. This type of ROI is useful if you would like to track objects across sections, as an example. Similarly to the polygon type, you have the option to "exclude" sections.
Input and Output: The ROI node has one input and one output. The input connects the ROI node with a previous node such as
video feed, and the output sends the results to the next node such as
Object Detection.
Related Articles
Overview Object Detection Node
The Object Detection Node is used to detect several different objects off the shelf with pre-trained or custom deep learning models on GPU, CPU, VPU (Intel Movidius Myriad X) or TPU (Google Coral). Input and Output Input: Frame from a video file, IP ...Overview Object Flow Node
The Object Flow Node detects and tracks people from an input video stream to compose a heatmap and to calculate the average dwell time. Input and Output Input: Object Detection mqtt result message, ROI section definition, Object Counting result ...Overview Object Tracking Node
The Object Tracking Node is used to track several objects which were previously detected by the object detection node across several frames. Input and Output Input: Output of object detection node. Output: MQTT message containing the tracking ...Overview Video Input Tools Module
The video input tools module includes nodes to grab the video frames and perform couple of pre-processing tasks on them. In particular, the video input tools include following nodes: Video Feed: The video feed node reads the input stream from the ...Overview Fall Detection Node
The Fall Detection Node detects and tracks the movement of people to identify if a person is falling. Input and Output Input: Group Keypoint Detection output message and stream Output: Fall detection message and stream Supported architecture: ...