The subject of this chapter is semantic alignment and deals with the conversion of multiple input data or measurements which do not refer to the same object, or phenomena, to a common object or phenomena. This is a core element in any data fusion algorithm since different inputs can only be fused together if the inputs refer to the same object or phenomena. In the chapter we first introduce the concept of an assignment matrix and its use in matching shapes and contours. Then we consider clustering algorithms considering in depth the K-means algorithm and its variants. This includes automatically estimating the number of clusters and robustifying the algorithm when outliers are present. In the last part of the chapter we consider cluster ensembles and how they may be fused together using the co-association matrix.