In this paper, we propose the use of Laplacian of Gaussian (LOG) filters in the detection of Cluster Mura defects and Vertical-Band Mura defects on Liquid Crystal Displays. To detect Cluster Mura defects, 2-D LOG filters with their shapes matching the shapes of Cluster Mura are adopted. The optimal threshold for the detection of Cluster Mura is determined based on the SEMU formula. On the other hand, to detect Vertical-Band Mura, a 1-D LOG filter is used over the projected 1-D intensity profile to check if the variation tendency of the intensity profile is changed. The portions with inconsistently intensity variation are regarded as the portions where V-Band Muras occur. The simulation results demonstrate the LOG filters are very useful in the detection of Mura defects.

Abstract— The defects of flat‐panel liquid‐crystal display are usually not uniform, and the defects are called as “Mura.” There are many factors that cause the Mura phenomenon. At present, the human eyes define the seriousness of Mura, and different operators define the same Mura phenomenon with different meanings. For this reason, there have been many conflicts between the suppliers of flat‐panel liquid‐crystal display and customers. In order to solve the conflicts between the suppliers and customers, some researchers proposed a regression equation for the luminance contrast threshold and the size of Mura. In addition to the effect of Mura size on luminance contrast threshold, this study investigated the relationship between other factors and luminance contrast threshold. An analysis of the results show that Mura size, polarity of Mura, Mura position, and different color backgrounds significantly affect the visual contrast threshold. However, polarity did not significantly affect visual contrast threshold. In this study, it was found that Mura size was more important than the other factors for visual contrast threshold. The results of this study is expected to be referenced for inspection by the LCD industry.

This paper describes a model of human spatial vision and a corresponding procedure for using the model to define a quality metric and Mura defects for flat-panel-displays. Together the proposed model and human contrast thresholds constitute a Mura quality metric which is responsive to noise generating both characteristics of a display system and a machine vision system. Predictions of the contrast perception performance of the model is presented and compared with human performance data. Results revealed that very small errors between predictions made by the model and the subjective test data. The results of the validation studies conducted so far suggest that the proposed method for Mura defect quantification is feasible and warrants critical examination.

Motion color images in LCDs become more and more important topic. In this paper, a color shift evaluation method is examined to judge the quality of captured motion images via a pursuit camera system. Furthermore, fifteen expectative color mixtures of color pairs were compared with results of subjective tests. Experiment results revealed that there exists high correlation between measuring system and the subjective test, and the proposed method was enough to be a quality index to present color shift characteristics for LCD displays.

In the present study, some factors were considered such as the various types and sizes of real Mura, and Mura inspection experience. The steps of data collection and experiments were conducted systematically from the viewpoint of human factors. From the experimental results, Mura size was the most important factor on visual contrast threshold. The purpose of this research was to objectively describe the relationships between the Mura characteristics and visual contrast thresholds. Furthermore, a domestic JND model of LCD industry was constructed. This model could be an inspection criterion for LCD industry.