# Object detection and display with LiteRT Source: [https://docs.qualcomm.com/doc/80-70022-50/topic/single-camera-stream-with-object-detection-and-display.html](https://docs.qualcomm.com/doc/80-70022-50/topic/single-camera-stream-with-object-detection-and-display.html) The use cases use a YOLOv5 LiteRT model to identify the object in a scene. The use case is to either overlay or compose the bounding boxes over the detected objects, and then display the results. ## Use qtivoverlay plugin to apply bounding box overlay Run the use case on the target device: gst-launch-1.0 -e --gst-debug=2 qtiqmmfsrc name=camsrc ! video/x-raw,format=NV12_Q08C,width=1280,height=720,framerate=30/1 ! \ queue ! tee name=split split. ! queue ! qtimetamux name=metamux ! queue ! qtivoverlay ! queue ! waylandsink fullscreen=true split. ! queue ! \ qtimlvconverter ! queue ! qtimltflite delegate=external external-delegate-path=libQnnTFLiteDelegate.so \ external-delegate-options="QNNExternalDelegate,backend_type=htp;" model=/etc/models/yolov5.tflite ! queue ! \ qtimlpostprocess settings="{\"confidence\": 75.0}" results=10 module=yolov5 labels=/etc/labels/yolov5.json ! text/x-raw ! queue ! metamux.Copy to clipboard To stop the use case, use CTRL + C. The following figure shows the flow of the use case execution: 1. Identifies object scenes in the scene from a video stream, which is coming through a camera source. 2. Overlays bounding boxes over the detected objects using overlaylib. 3. Displays the results. Figure : Pipeline for bounding box overlay Qualcomm Open source camsrc tee qtimlvconverter qtimltflite qtimlpostprocess qtimetamux qtioverlay waylandsink The following table provides the sequential processing stages of the pipeline execution: Table : Pipeline processing stages for bounding box overlay | Process | Description | | --- | --- | | qtiqmmfsrc |

  1. Collects the video stream (source) and creates two copies of
    the source:


| | **Preprocessing** | | qtimlvconverter |

  1. Receives the video stream on its sink pad.


  2. Performs preprocessing:


  3. Converts the video stream to a tensor stream on its source
    pad.

    The object detection model uses this tensor
    stream for inferencing.




| | **Inferencing** | | qtimltflite |

  1. Loads the object detection model.


  2. Modifies the graph for the chosen delegate.


  3. Receives the tensor stream on its sinkpad.


  4. Runs the inference and produces a tensor stream with the
    object detection results on its source pad.


| | **Postprocessing** | | qtimlpostprocess |

  1. Receives the inference tensors from object detection.


  2. Converts the inference tensors on its sinkpad into formats
    like video or text that the multimedia plugins can process
    later.


  3. Applies the threshold to the chosen number of results.


  4. Loads the corresponding modules for detection models.

    In
    this use case, qtimlpostprocess does the following:


    1. Loads the YOLOv5 submodule.


    2. Produces results as structures of text.


    3. Sends them to the sinkpad of qtimetamux.





| | qtimetamux |

  1. Receives video stream and text stream with bounding box
    results corresponding to the video stream on its
    sinkpads.


  2. Produces GST buffers with contents of video stream from its
    sink pad.


  3. Adds bounding boxes as GstVideoRegionOfInterest from data
    sinkpad to GST buffers meta (meta muxing) on its source
    pad.


| | qtivoverlay |

  1. Receives the multiplexed stream.


  2. Overlays the bounding boxes on the VideoFrame using CL.


  3. Produces GST buffers with overlays in its source pad.


| | **Output** | | Waylandsink |

  1. Receives the video stream on its sinkpad.


  2. Submits the video stream to Weston.


  3. Weston renders the video stream and bounding boxes generated
    for the objects in that scene on a local display
    device.


| ## Use qtivcomposer to mix original frame with bounding box mask Run the use case on the target device: gst-launch-1.0 -e --gst-debug=2 qtiqmmfsrc name=camsrc ! video/x-raw,format=NV12_Q08C,width=1280,height=720,framerate=30/1 ! queue ! tee name=split split. ! \ queue ! qtivcomposer name=mixer sink_1::dimensions="<1920,1080>" ! queue ! waylandsink fullscreen=true split. ! queue ! qtimlvconverter ! queue ! \ qtimltflite delegate=external external-delegate-path=libQnnTFLiteDelegate.so external-delegate-options="QNNExternalDelegate,backend_type=htp;" \ model=/etc/models/yolov5.tflite ! queue ! qtimlpostprocess settings="{\"confidence\": 75.0}" results=10 module=yolov5 labels=/etc/labels/yolov5.json \ ! video/x-raw,format=BGRA,width=640,height=360 ! queue ! mixer.Copy to clipboard To stop the use case, use CTRL + C. The following figure shows the flow of the use case execution: 1. Identifies object scenes in the scene from a video stream, which is coming through a camera source. 2. Composes the following using qtivcomposer: 1. Bounding boxes over objects detected. 2. Original video stream. 3. Displays the results. Figure : Pipeline for bounding box mask with qtivcomposer Qualcomm Open source camsrc tee qtimlvconverter qtimltflite qtimlpostprocess qtivcomposer waylandsink The following table provides the sequential processing stages of the pipeline execution: Table : Pipeline processing stages for bounding box mask with qtivcomposer | Process | Description | | --- | --- | | qtiqmmfsrc |

  1. Collects the video stream (source) and creates two copies of
    the source:


| | **Preprocessing** | | qtimlvconverter |

  1. Receives the video stream on its sink pad.


  2. Performs preprocessing:


  3. Converts the video stream to a tensor stream on its source
    pad.

    The object detection model uses this tensor
    stream for inferencing.




| | **Inferencing** | | qtimltflite |

  1. Loads the object detection model.


  2. Modifies the graph for the chosen delegate.


  3. Receives the tensor stream on its sinkpad.


  4. Runs the inference and produces a tensor stream with the
    object detection results on its source pad.


| | **Postprocessing** | | qtimlpostprocess |

  1. Receives the inference tensors from the object detection
    model.


  2. Converts the inference tensors on its sinkpad into formats
    like video or text that the multimedia plugins can process
    later.


  3. Applies the threshold to the chosen number of results.


  4. Loads the corresponding modules for detection models.

    In
    this use case, qtimlpostprocess does the following:


    1. Loads the YOLOv5 submodule.


    2. Produces video frames with only bounding boxes that
      can be overlaid on objects.


    3. Sends them to the sinkpad of qtivcomposer.





| | qtivcomposer |

  1. Receives the original video stream and video stream with
    bounding boxes on its sinkpads.


  2. On its sourcepads, produces content that's composed of the
    video streams processed from its sinkpads.


| | **Output** | | Waylandsink |

  1. Receives the video stream on its sinkpad.


  2. Submits the video stream to Weston.


  3. Weston displays the following on a local display device:

    1. The video stream is captured from the camera.


    2. The bounding boxes are drawn over the allowed number
      of objects identified in that scene.





| **Parent Topic:** LiteRT use cases Last Published: Feb 20, 2026 Previous Topic Audio classification decode and display with LiteRT Next Topic Object detection and encode with LiteRT