FlowTSE: Target Speaker Extraction with Flow Matching