EDGEMATRIX Stream Toolkit Tutorial

Contents:

• Overview
  • 1. The concept of Edge AI
  • 2. EDGEMATRIX Service
    • Toolkit Box
  • 3. EDGEMATRIX Stream
    • EDGEMATRIX Stream Application Package
    • EDGEMATRIX Stream Controller
    • EDGEMATRIX Stream Pipeline
  • 4. The development process using the EDGEMATRIX Toolkit
• Quick Start
  • Update to the latest version
    • Check the current version
    • Run an update script
  • Prepare a template app
  • Create a new EAP by copying from a template EAP
    • runtoolkit and toolkit_home
    • Launch the EDGEMATRIX Stream Toolkit application
    • Create a new EAP
    • Select a new EAP
  • Validate the new EAP
    • Open a validation dialog
    • Run a validation
    • Use your own sample siginal to validate
  • Test the new EAP
    • Execute, Choose a stream and Create an EAP package
    • Play a pipeline
    • Stop a pipeline
    • Movie files made by record actions
• Tutorials
  • Preparing a simple Detector project
  • Using your own events and callback function
    • Write your own callback function
    • Use your own event
    • Test your application
  • Using your own input source
    • Preparing your movie file for a streaming
  • Using your own trained Yolo model binary with IPlugin
    • Follow the Yolo tutorial on the DeepStream SDK
    • Place your trained Yolo model binary and related files
    • Change property configurations
  • Making a submission package
    • What is a submission package?
    • How to make a submission package?
• Sample App - Parking Violation Detector
  • The problem that this sample app solves
  • Design Parking Violation Detector
    • What actions should be taken under what conditions?
    • How to let an end user to customize an action?
    • How to let an end user to define a restricted area?
  • Prepare your pipeline
    • AI model binaries and engine files
    • Other elements
    • DeepStream config
    • Test the pipeline
    • Template app
  • Write a callback
    • Define the structure of an event
    • Generate an event
    • Define Polygons in Options
    • Implementation
  • Action Definitions
    • SNMP Action
    • LINE Action
    • Recording Action
    • Upload Action
• Configurations
  • Toolkit Directory Structure
  • Application Configurations
    • Overview
    • Input
    • Primary/Secondary
    • Meta Transfer Mode
    • Tracker
    • EMCustom
    • EMPyCustom
    • Overlay
    • AI Meta
    • Callback and Events
    • Options
  • Stream Configurations
    • Actions
    • Continuous Recording
• Templates
  • DeepStream samples
    • EMI Pedestrian DCF Counter
    • EMI Pedestrian IOU Counter
    • EMI Pedestrian KLT Counter
    • EMI Vehicle Counter
    • EMI Vehicle DCF Counter By Color
    • EMI Vehicle IOU Counter By Make
    • EMI Vehicle Multiple Secondaries
    • Yolo V3 Detector
    • Yolo V3 Tiny Detector
    • EMI Semantic Segmentation
  • Custom templates
    • EMI Vehicle EMCustom
    • EMI Vehicle Color Watcher
    • EMI EMCustom CPP
    • EMI Vehicle PrePost EMCustom
    • EMI Torch 3DCNN
    • EMI Torch Imagenet Yolo IOU Counter
    • Facenet
  • IoT templates
    • EMI IoT Sensor Monitor
    • EMI Simple UDP
    • EMI Simple Recorder
• Double EAPs
  • Description
    • 1. To add a Secondary only EAP
    • 2 To use two cameras (RTSP streams)
  • Examples
    • Preparation
    • Create Apps
    • 1. To add a Secondary only EAP
    • 2. To use two cameras (RTSP streams)
• EMCustom
  • Description
  • Properties and Signals
  • Usage and Examples
  • Debugging
    • Visualizing output frame
    • Print output meta to console
    • Reading the output meta from the property
    • Wait for signal before reading last meta property
  • Interface
  • How to add a custom library
  • EMCustom Meta
    • Array structure
    • Object structure
  • EMCustom Integration
  • Use Dynamic Libraries With EMCustom
  • Access metadata structures With EMCustom
  • Examples
    • Passthrough
    • Average Intensity in a person ROI
    • EMCustom options
    • Batch meta from custom library
• EMPyCustom
  • Description
  • Properties and Signals
  • Element Caps
  • Virtual Methods
  • Usage and Examples
    • Mapping the buffer
    • Example pipeline
  • Debugging
    • Visualizing output frame
    • Print output meta to console
  • How to add a custom library
  • EMPyCustom Meta
  • EMPyCustom Integration
  • Examples
    • Average Intensity in a person ROI
    • Torch classification CPU
  • EmPyCustom options
  • Known issues
    • Gst-python
    • DeepStream python
• Custom AI pipeline
  • Overview
  • Supported Elements
  • Considerations
  • Possible Combinations
    • Basic Patterns without EMPyCustom
    • Advanced Patterns with EMPyCustom
• Sample Torch Apps
  • Just Torch
    • Why just Torch?
    • Development/Runtime Environment
  • Torch reference apps
  • EMI Torch 3DCNN
    • Pipeline Configuration
    • Torch Model
    • Torch Program
    • Stream Configuration
    • Execution
  • EMI Torch Imagenet Yolo IOU Counter
    • Pipeline Configuration
    • Torch Model
    • Torch Program
    • Stream Configuration
    • Execution
• EDGEMATRIX Stream CLI
  • edgestream_app
  • EDGEMATRIX Stream State Diagram
  • CLI Commands
    • start_webrtc
  • CLI Result Messages
    • Command result message
    • Result message available
    • Result Error Messages
• Known Issues
  • v2.3
    • Known Memory Leaks
  • v1.6.2
    • Pre processing of a primary nvinfer of a custom pipeline prevents it from running
  • v1.1
    • A record action could hang a GStreamer daemon process if a debug window is shown