Abstract: Images of semiconductor wafers can be hashed to determine a fixed length hash string for each of the images. Pattern synonyms can be determined from the hash strings. The pattern synonyms can be grouped. A degree of similarity between images in the groups is adjustable via a hamming distance. This can be used for various applications, including determination of latent defects.

Abstract: Images of semiconductor wafers can be hashed to determine a fixed length hash string for each of the images. Pattern synonyms can be determined from the hash strings. The pattern synonyms can be grouped. A degree of similarity between images in the groups is adjustable via a hamming distance. This can be used for various applications, including determination of latent defects.

Abstract: Embodiments herein include methods, systems, and apparatuses for die screening using inline defect information. Such embodiments may include receiving a plurality of defects, receiving wafersort electrical data for a plurality of dies, classifying each of the defects as a defect-of-interest or nuisance, determining a defect-of-interest confidence for each of the defects-of-interest, determining a die return index for each of the dies containing at least one of the defects-of-interest, determining a die return index cutline, and generating an inking map. Each of the defects may be associated with a die in the plurality of dies. Each of the dies may be tagged as passing a wafersort electrical test or failing the wafersort electrical test. Classifying each of the defects as a defect-of-interest or nuisance may be accomplished using a defect classification model, which may include machine learning. The inking map may be electronically communicated to an inking system.

Abstract: Embodiments herein include methods, systems, and apparatuses for die screening using inline defect information. Such embodiments may include receiving a plurality of defects, receiving wafersort electrical data for a plurality of dies, classifying each of the defects as a defect-of-interest or nuisance, determining a defect-of-interest confidence for each of the defects-of-interest, determining a die return index for each of the dies containing at least one of the defects-of-interest, determining a die return index cutline, and generating an inking map. Each of the defects may be associated with a die in the plurality of dies. Each of the dies may be tagged as passing a wafersort electrical test or failing the wafersort electrical test. Classifying each of the defects as a defect-of-interest or nuisance may be accomplished using a defect classification model, which may include machine learning. The inking map may be electronically communicated to an inking system.