A comprehensive, research-backed course on real-time object detection. You’ll go from understanding the theoretical foundations of convolutional neural networks and anchor-based architectures to deploying optimized YOLO, RT-DETR, and custom detection models at production scale.

What you’ll master:
• CNN architectures for feature extraction (ResNet, CSPDarknet, EfficientNet)
• YOLO family deep-dive (YOLOv5 through YOLOv10, architecture choices & tradeoffs)
• Transformer-based detectors (DETR, RT-DETR) and when to use them
• Training pipelines: data augmentation, mosaic, mixup, hyperparameter tuning
• Custom dataset preparation, annotation best practices, and label quality
• Model optimization: quantization, pruning, TensorRT, ONNX export
• Edge deployment on Jetson, Raspberry Pi, and mobile (CoreML, TFLite)
• Evaluation metrics: mAP, IoU, precision-recall curves, COCO benchmarks
• Live inference pipelines with OpenCV, GStreamer, and async video processing

Taught by an NLP researcher who crossed into computer vision — you’ll get the rigorous, paper-reading approach combined with practical engineering.