MATLAB Implementation of Curvature-Based Method for Finger Detection

To achieve more accurate finger position detection, we implemented a curvature-based method in MATLAB. The algorithm calculates curvature by computing angles formed by three equally spaced points along the finger contour - a computationally efficient approach that uses vector dot products and trigonometric functions. The core implementation involves sliding a three-point window along the contour and applying thresholding to distinguish between high-curvature fingertip regions and lower-curvature finger root areas. This method not only improves detection accuracy by reducing errors but also incorporates algorithmic optimizations for handling diverse hand gestures, including dynamic threshold adjustment based on contour length normalization. The system's stability and reliability are enhanced through robust edge-case handling in the MATLAB code, such as managing contour endpoints and smoothing irregular curves. Overall, this straightforward yet effective approach provides a solid foundation for subsequent gesture recognition and human-computer interaction applications, with the MATLAB implementation featuring clear modular functions for curvature calculation, threshold processing, and region classification.