Learning Vector Quantization Neural Networks for LED Wafer Defect Inspection

A case study of applying regional level-set formulation to post-sawing LED wafer inspection

“…In contrast, the proposed approach uses only two steps while achieving higher integral accuracy than that (93.5%) in [4]. Hence, the proposed approach is suitable for other cases with only minor reconfiguration.…”

“…In [4], die extraction, preprocess filtering, light compensa tion, segmentation, feature extraction, and inspection were required. In contrast, the proposed approach uses only two steps while achieving higher integral accuracy than that (93.5%) in [4].…”

“…Chang et al developed an inspection system that looks at various functional parts of the post-sawing dies [4]. They utilized learning vector quantization neural networks to in spect the parts using geometric and textural features.…”

“…In this study, the evolution of the zero-level set is applied to wafer defect inspection. Unlike existing methods [1][2][3][4], the pro posed approach requires only two steps -segmentation and inspection.…”

“…Many inspection methods have been proposed [1][2][3][4]. Tobin et al proposed strategies for integrated yield management based on automatic defect classification, spatial signature analysis, and automatic image retrieval technologies [1].…”