Eyeglasses detection based on learning and non-learning based classification schemes

In this paper, we propose two schemes for prescription eyeglasses detection. The first proposed scheme is non-learning which uses Viola-Jones to detect Region of Interest (ROI) followed by glass detection yielding an overall accuracy of 99.0% for FERET and 97.9% for VISOB datasets. The second scheme is the learning-based scheme consisting of three main steps (a) ROI detection, (b) Local Binary Pattern (LBP) and Histogram of Gradients (HOG) feature extraction and fusion them together, and (c) applying classifiers such as Support Vector Machine (SVM), Multi-Layer Perceptron (MLP), and Linear Discriminant Analysis (LDA), and fusing the output of these classifiers. The latter obtained a best overall accuracy of about 99.3% on FERET and 100% on VISOB dataset.