Hand Recognition Using a Fuzzy Classifier

Wen-Yen Wu

hand recognition; fuzzy Classifier; compactness; feature

This paper proposes a hand recognition method using a fuzzy k nearest neighbor Classifier. Firstly, it obtains the dominant points, and calculate the perimeter of the triangle formed by two consecutive dominant points and the center of gravity of the polygon, and the compactness of the included angle of the triangle formed by three consecutive dominant points, etc. as feature values. Then use the fuzzy k-neighbor cluster analysis method to properly classify the eigenvalues, and then calculate the center of gravity of all the eigenvalues of each category as the representative eigenvalues of the category, so that the number of samples of the eigenvalues tested can be reduced. It can also identify the hands more conveniently and quickly, saving identification time. Some experiments have been conducted and the experimental results indicate that the proposed method can recognize the hand shapes effectively.

https://doi.org/10.31871/IJNTR.8.12.3