Rotation Invariant Face Detection Using Wavelet, PCA and Radial Basis Function Networks

TL;DR

This paper presents a new face detection method that uses wavelet analysis, PCA for dimensionality reduction, and RBF neural networks to identify faces and their orientations with improved accuracy and robustness.

Contribution

The paper introduces a novel combination of wavelet, PCA, and RBF networks for rotation-invariant face detection, outperforming traditional RBF approaches.

Findings

Better approximation capability than traditional RBF networks

Faster learning speed and smaller network size

High robustness to outliers

Abstract

This paper introduces a novel method for human face detection with its orientation by using wavelet, principle component analysis (PCA) and redial basis networks. The input image is analyzed by two-dimensional wavelet and a two-dimensional stationary wavelet. The common goals concern are the image clearance and simplification, which are parts of de-noising or compression. We applied an effective procedure to reduce the dimension of the input vectors using PCA. Radial Basis Function (RBF) neural network is then used as a function approximation network to detect where either the input image is contained a face or not and if there is a face exists then tell about its orientation. We will show how RBF can perform well then back-propagation algorithm and give some solution for better regularization of the RBF (GRNN) network. Compared with traditional RBF networks, the proposed network demonstrates better capability of approximation to underlying functions, faster learning speed, better size of network, and high robustness to outliers.