Quantitative reconstruction of objects from spatially correlated image sequences

Multivariate image analysis (MIA) is classically used on image sequences, in order to exhibit inter-images correlation. The goal of this study is to demonstrate that it can also be used on correlated image sequences. Such dataset can be obtained when using multi-sensor technique such as secondary ion mass spectrometry (SIMS). It is then possible to compute analyte relative abundance from principal component loadings and reconstruct complex sample internal structure. The principles of the technique are presented and applied to multi-layer electronic component. It was possible to show that the sample was not correctly manufactured and that several unexpected layers have been added.