object-detection-opencv

Why we included this project

This repo shows how to run YOLO v3 object detection through OpenCV's dnn module using just numpy and opencv-python. You supply an image, config file, weights, and class list via a single command line, and it draws the bounding boxes. The code is a few short scripts, simple enough to read through in one sitting and modify for your own experiments. Teams testing whether OpenCV's built-in inference fits a quick prototype will find the command-line setup easy to copy. It is a tutorial example rather than a deployable service, but that makes it a solid way to try detection locally without a heavy deep-learning stack.