Object detection is no longer a research problem — it is a deployment problem. YOLOv8 runs at video frame rates on laptops, Raspberry Pis, and browser tabs. This book builds the pipeline around it. Start with images as data: NumPy arrays, BGR channels, color spaces, filtering. Move into OpenCV's DNN module — loading ONNX models, parsing the YOLOv8 output tensor, running non-maximum suppression. Then build real-time video pipelines, train YOLO on your own dataset, track objects with ByteTrack, optimize with FP16 and INT8, and deploy to the edge.Working code throughout. Runs on any machine.
