Reaves.dev

v0.1.0

built using

Phoenix v1.7.12

Persistent Temporal Streams

Stephen M. Reaves

::

2024-04-27

Notes about Lecture 10b for CS-6210

Summary

Programming Paradigms

Novel Multimedia Apps

Sensor-based

Sensors -> distribution -> computation

Programming Model for Situation Awareness

Sequential Program for video analytics

GCameraCameraDetectionDetectionCamera->DetectionTrackingTrackingDetection->TrackingRecognitionRecognitionTracking->RecognitionAlarmAlarmRecognition->Alarm

Objective in Situation Awarenes apps

How do we scale?

PTS Programming Model

Threads talk to other threads through channels to create a process graph

Channels holds sequenced data objects ordered by timestamps

Bundling Streams

Gcluster_anchoranchorcluster_dependantdependants1VideoVideos1->Videos2AudioAudios2->Audios3TextTexts3->Texts4GestureGestures4->Gestures5Video->s5Audio->s5Text->s5Gesture->s5

Can name stream groups and name an anchor

groupget: Get corresponding time-stamped streams from all streams in the group

Power of Simplicity

Sequential Program for video analytics

GCameraCameraDetectionDetectionCamera->DetectionTrackingTrackingDetection->TrackingRecognitionRecognitionTracking->RecognitionAlarmAlarmRecognition->Alarm

Distributed Program for video analytics

GCaptureCaptureframesframesCapture->framesDetectionDetectionblobsblobsDetection->blobsTrackingTrackingobjectsobjectsTracking->objectsRecognitionRecognitioneventseventsRecognition->eventsAlarmAlarmframes->Detectionblobs->Trackingobjects->Recognitionevents->Alarm

PTS Design Principles

pts channel arch