casdome.blogg.se

How to set up stream labels
How to set up stream labels






how to set up stream labels

Track awareness is specified in the source stream service. When a stream layer is track aware, you can symbolize the current (most recent) observations one way and the previous (older) observations another way. Symbolize track-aware stream layersĪ track-aware stream layer relies on an attribute to group together observations that correspond to a single object. A Cartographic offset can be configured on the Elevation tab of the Layer Properties dialog box for the stream layer. To alleviate this, consider adding a small cartographic offset to stream layers in this situation. Polygon and polyline stream layers may visually conflict with the ground surface in a scene, making them only partially visible. They cannot be draped layers in the 2D category. You input start and stop information to determine the processing time.Stream layers can only be placed in the 3D Layers category of a scene. StartStreamProcessor operation on a label detection stream processor, You start analyzing streaming video by calling StartStreamProcessor with the stream processor name that you specified Starting the Amazon Rekognition Video label detection stream processor Then you can set MinConfidence to be lower than 50 to receive more notifications. If you want to optimize the algorithm for higher recall,

how to set up stream labels

You then receive fewer notification, but each notification is more reliable. If you want to optimize the algorithm for higher precision, then you can set MinConfidence

how to set up stream labels

For example, if you want to be notified only when Rekognition is absolutely certain there is a package in the video frameīy default, MinConfidence is set to 50. You can set MinConfidence to a desired value of your choice between 0 and 100ĭepending on how frequently you want to be notified. MinConfidence is a numerical value ranging from 0 to 100 that indicates how certain the algorithm is about its predictions.įor example, a notification for person with a confidence value of 90 means that the algorithm is absolutely certain that the person is present in the video.Ī confidence value of 10 indicates that there might be a person. Note that you can change the MinConfidence value when you specify the ConnectedHomeSettings for the stream processor. "RoleArn": "arn:aws:iam::nnnnnnnnn:role/Admin", "SNSTopicArn": "arn:aws:sns:us-east-2:nnnnnnnnnnnn:MyTopic" "Arn": "arn:aws:kinesisvideo:us-east-1:nnnnnnnnnnnn:stream/muh_video_stream/nnnnnnnnnnnnn"








How to set up stream labels