WK.

YOLOv8 Dog Pose Estimation

An end-to-end Animal Pose Estimation project focused on dog keypoint detection using YOLOv8. The system is split into two main parts: a training pipeline integrated with Roboflow for dataset management, and an interactive Web Inference application built with Gradio for real-time testing on images, videos, and webcams.

YOLOv8 Dog Pose Estimation - 1
1 / 12

My Role

Computer Vision Engineer — Dataset Preparation, Model Training, Web Interface Development

Tech Stack

PythonYOLOv8Computer VisionGradioPyTorchOpenCVRoboflow

Source Code

wongsakron/yolov8-pose-dog

Key Highlights

  • Custom-trained YOLOv8 Pose model specifically for dog keypoint detection
  • Integration with Roboflow API for streamlined dataset and weights management
  • Interactive web interface built with Gradio for real-time model inference
  • Support for analyzing static images, video files, and live webcam feeds