YOLO Object Detection and Segmentation with Python

This 3-week course offers a practical and beginner-friendly introduction to object detection and segmentation using modern YOLO models. Participants will start with pretrained models trained on large-scale datasets such as ImageNet and progressively build the skills needed to prepare custom datasets, train and evaluate YOLO models, and apply them in real-world scenarios. Applications include identifying animals in smart agriculture, detecting and segmenting lesions in medical images, monitoring safety gear in industrial settings, counting vehicles in traffic surveillance, and deploying models to web apps, APIs, or edge devices like Raspberry Pi or Jetson Nano. The focus is on hands-on learning, with each lesson supported by a ready-to-use Jupyter notebook. Week 1 – Getting Started with YOLO Participants are introduced to the basics of object detection, install modern YOLO models, and run inference on images, videos, and webcam streams. They explore how to interpret model outputs and enhance results through image preprocessing and basic augmentation techniques. Week 2 – Training a Custom YOLO Model Participants annotate their own datasets using tools like Roboflow or CVAT, configure training using YOLO’s YAML format, and evaluate model performance with key metrics such as mAP and confusion matrices. They also learn how to tune hyperparameters and apply augmentations to improve training outcomes. Week 3 – Advanced Applications and Deployment In the final week, participants explore object segmentation, real-time video processing, and build simple web and API applications using Streamlit and FastAPI. They also learn how to process large batches of images and deploy trained models to edge devices for offline use.