To investigate the impact of melt conditioning and filtration on intermetallic phases in AlSi9Cu3 µCT studies were conducted. The gained research data show microstructural features like pores (#1) and polyhedral intermetallic phases (#2, #2) in 3D videos. The fatigue crack introduced by ultrasonic fatigue testing (USFT) was segmented from µCT slices by machine learning algorithms of Trainable Weka Segmentation and superimposed with the initial microstructure. The superposition of fatigue crack and initial microstructure is shown for a reference batch specimen (#4) and a melt con-ditioned and filtrated batch specimen (#5).