YOLOv12: Custom Object Detection, Tracking & WebApps | Udemy

YOLOv12 is the latest state-of-the-art computer vision model architecture, surpassing previous versions in both speed and accuracy. Built upon the advancements of earlier YOLO models, YOLOv12 introduces significant architectural and training enhancements, making it a versatile tool for various computer vision tasks.

The YOLOv12 model supports a wide range of tasks, including object detection, instance segmentation, image classification, pose estimation, and oriented object detection (OBB).

Course Structure

This course is divided into multiple sections, covering everything from the fundamentals of YOLOv12 to advanced applications.

Introduction to YOLOv12

• What’s New in YOLOv12
• Key updates and features in YOLOv12
• Non-Maximum Suppression & Mean Average Precision in Computer Vision

Running YOLOv12

• Setting up YOLOv12
• Using YOLOv12 for Object Detection
• Evaluating YOLOv12 Model Performance: Testing and Analysis

Dataset Preparation

• How to find and prepare datasets
• Data annotation, labeling, and automatic dataset splitting

Training YOLOv12

• Fine-Tuning YOLOv12 for Object Detection on Custom Datasets
• Custom Projects:
  • Train YOLOv12 for Personal Protective Equipment (PPE) Detection
  • Train YOLOv12 for Potholes Detection

Advanced Multi-Object Tracking

• Implementing Multi-Object tracking with Bot-SORT and ByteTrack algorithms

Advanced Applications

• Blurring Objects with YOLOv12 and OpenCV-Python
• Generating Intensity Heatmaps to Identify Congestion Zones
• Building a Tennis Analysis System with YOLO, OpenCV, and PyTorch

Web Integration

• Developing Web Apps with YOLOv12 and Flask