Skip to content

Latest commit

 

History

History
56 lines (37 loc) · 1.88 KB

tiny_yolo.rst

File metadata and controls

56 lines (37 loc) · 1.88 KB
Raw

RGB & Tiny YOLO

This example shows how to run YOLO on the RGB input frame, and how to display both the RGB preview and the metadata results from the YOLO model on the preview. Decoding is done on the RVC instead on the host computer.

Configurable, network dependent parameters are required for correct decoding:

  • setNumClasses - number of YOLO classes
  • setCoordinateSize - size of coordinate
  • setAnchors - yolo anchors
  • setAnchorMasks - anchorMasks26, anchorMasks13 (anchorMasks52 - additionally for full YOLOv4)
  • setIouThreshold - intersection over union threshold
  • setConfidenceThreshold - confidence threshold above which objects are detected

By default, Tiny YOLOv4 is used. You can add yolo3 as a CMD argument to use Tiny YOLOv3.

Demo

Setup

Source code

Python

Also available on GitHub

../../../../examples/Yolo/tiny_yolo.py

C++

Also available on GitHub

../../../../depthai-core/examples/Yolo/tiny_yolo.cpp