PROJECT TITLE :

Ambiguous Surface Defect Image Classification of AMOLED Displays in Smartphones

ABSTRACT:

During this paper, we tend to propose a classification approach for ambiguously formed defects found on the surface of a sort of show panel module that is widely employed in the field of mobile displays. These sorts of surface defects are difficult to properly distinguish due to defect similarity and variety. In such cases, defect sorts can only be determined using cumbersome human visual inspection. To solve the matter of ambiguous surface defect classification, we tend to introduce a unique filtering methodology that effectively separates the foreground defective regions from the background, that has structured patterns, local illumination variation, and completely different light-weight conditions for each of several cameras in an inspection system. Applying the proposed filter methodology to defect pictures, we select necessary options by adopting a wrapper-based mostly feature choice technique employing a random forest as a learning algorithm. Successful classification results using the presented model are obtained using difficult real-world defect image knowledge gathered from a smart phone display module inspection line in an industrial plant.