=================================================================================

A semiconductor wafer includes many dies. The detection and classification of defective dies (die-scale inspection) in wafer map can be performed by:
         i) Template-matching based methods (traditional).
         ii) Automated visual inspection (AVI) with template extraction algorithm and then a deep learning-based classification: [1]
            ii.a) Standard template extraction to predict clustering by arranging the columns and rows in the detected dies.
              ii.a.1) Abtain a standard template from the original scanning acoustic tomography (SAT) image.
              ii.a.2) A majority of the die patterns are detected through template matching.
            ii.b) Die detection through template matching.
            ii.c) Pattern candidate prediction through clustering.
            ii.d) Pattern classification through deep learning.

============================================

[1] Hsiang-Chieh Chen, Automated Detection and Classification of Defective and Abnormal Dies in Wafer Images, Appl. Sci. 2020, 10, 3423; doi:10.3390/app10103423, (2020).