Automatic Defect Classification of TFT-LCD Panels with Shape, Histogram and Color Features

In this paper, an automatic defect classification algorithm for thin film transistor liquid crystal display (TFT-LCD) manufacturing is proposed. Each sample of defect data contains three images: the original image, the defect shape image and the circuit zone image. A set of features including shape, histogram and color is extracted. Some common classifiers were tested in the experiments and Linear-SVM (Linear Surport Vector Machine) was chosen in practical manufacturing. A novel LBP-E feature considering intensity equality proposed in this paper is compared to other original rotation invariant LBP (Local Binary Pattern) features. The experimental results show that our method can generate a better result with a relatively low dimension number.