Dead Pixel Detection on Liquid Crystal Displays using Random Forest, SVM, and Harris Detector

Manufacturing TVs and monitors requires an effective method to detect the dead pixels. These defects are usually identified by operators manually. However, manual inspection is susceptible to failure due to human fatigue and generate a high cost for production process. In this work, conducted by three partners (UFAM/CETELI, ICTS and ENVISION/TPV), we propose three methods for automated detection of dead pixels. Two proposed methods were based on machine learning (ML) techniques, random forest (RF) and support vector machine (SVM) algorithms. The third method was based in digital image processing (DIP), Harris algorithm. As result, the SVM obtained the better performance with 92.1% of precision in two of three used image database.