The importance of statistical lifetime models for the reliability assessment of semiconductors is increasing steadily, because resources and time are limited. The devices are tested under accelerated electrical and thermal conditions which causes degradation in metal layers. To visualize the damage, Scanning Acoustic Microscopy (SAM) is used. In this work, an approach combining image processing and statistical modeling is presented in order to quantify and predict the damage intensity in SAM images. The image processing algorithm automatically locates and quantifies the maximum damaged areas in SAM images. The damage intensity is coded as an ordered categorical variable and a cumulative link model for damage prediction is defined. Both the algorithm and the proposed statistical model show good results.
CITATION STYLE
Alagić, D., Bluder, O., & Pilz, J. (2018). Quantification and Prediction of Damage in SAM Images of Semiconductor Devices. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10882 LNCS, pp. 490–496). Springer Verlag. https://doi.org/10.1007/978-3-319-93000-8_55
Mendeley helps you to discover research relevant for your work.