Head tracking using OpenCV
This Godot module demonstrates head tracking using OpenCV and CUDA for improved performance. It adjusts the camera's perspective in a 3D environment based on the user's head movements captured via webcam.
https://github.com/antarktikali/godot-opencv-gpu-perspectiveHead Tracking with OpenCV in Godot
This project provides a Godot Engine module that utilizes OpenCV to implement head tracking. By using a webcam, the module detects the user's face and adjusts the camera's position and angle within the 3D environment, creating a sense of perspective. This approach aims to enhance the user's immersion and interaction with the game world.
GPU-Accelerated Face Detection
Traditional CPU-based face detection can be slow, leading to lag and a poor user experience. This module addresses this issue by leveraging OpenCV's GPU module and CUDA to accelerate face detection using Haar cascades. While not perfect, this approach significantly improves performance compared to CPU-based methods.
Dependencies and Setup
To use this module, you'll need a recent version of OpenCV compiled with GPU support and a CUDA-compatible GPU. The module is designed to be placed in the modules directory of your Godot engine installation. After compiling Godot, you can open the example project provided to explore the implementation. The path to the Haar cascade XML file needs to be specified correctly in the OpenCVPerspective.cpp file.
Usage
When the start() function is called, a new thread is created to connect to the webcam and process frames for face detection. The get_horizontal_offset() and get_vertical_offset() functions allow you to retrieve the head's offset from the center of the screen. These offset values can then be used to manipulate the camera's transform values, achieving the desired head tracking effect.