AI Inference Conversion – YOLOv8 & Pose Estimation (Python ➡ C++)

This project demonstrates how to convert minimal AI inference scripts written in Python to equivalent C++ implementation using ONNX, OpenCV, and other inference engines.

---

🔍 Task Overview

Task 1: YOLOv8 Object Detection

• Original Language: Python
• Converted To: C++
• Model Used: YOLOv8n
• Inference Framework: ONNX Runtime / OpenCV DNN
• Input: person.jpg
• Output: Detected bounding boxes overlaid on image

Task 2: Human Pose Estimation

• Original Language: Python (Hugging Face model: microsoft/posex)
• Converted To: C++
• Note: Due to unavailability of a working ONNX version of microsoft/posex, a COCO-based OpenPose alternative was used.
• Inference Framework: OpenCV / ONNX
• Input: person.jpg
• Output: Human keypoints (poses) visualized

Requirements

• C++ Compiler (e.g., g++)
• OpenCV with DNN module
• ONNX Runtime (depending on code path)