This section includes example configurations that demonstrate how to configure feedback chaining.
The following configuration, for the upstream Media Server, runs face detection and then sends records to a Media Server at gpu-mediaserver:14000
for remote face recognition:
[Ingest] IngestEngine=AV [AV] Type=LibAV [Analysis] AnalysisEngine0=FaceDetect AnalysisEngine1=RemoteAnalysis [FaceDetect] Type=FaceDetect FaceDirection=Front MinSize=200 SizeUnit=pixel [RemoteAnalysis] Type=RemoteAnalysis Host=gpu-mediaserver Port=14000 ConfigName=RemoteFaceRecognition Input=DetectedFaces:Crop.Output Output=RecognizedFaces:FaceRecognition.Result [Transform] TransformEngine0=Crop [Crop] Type=Crop Input=FaceDetect.ResultWithSource [Output] OutputEngine0=XML [XML] Type=XML Input=RemoteAnalysis.RecognizedFaces XMLOutputPath=./output/html/%segment.type%_results_%segment.sequence%.html XSLTemplate=./xsl/tohtml.xsl Mode=Time OutputInterval=30s
The following configuration, for the remote Media Server, runs face recognition on records received from the upstream Media Server. To match the upstream configuration, above, this should be saved as RemoteFaceRecognition.cfg
, in the folder specified by the ConfigDirectory
parameter on the remote Media Server.
[Ingest] IngestEngine=RecordsFromUpstream [RecordsFromUpstream] Type=Receive Input=DetectedFaces [Analysis] AnalysisEngine0=FaceRecognition [FaceRecognition] Type=FaceRecognize Input=RecordsFromUpstream.DetectedFaces RecognitionThreshold=60 MaxRecognitionResults=1
|