Reliable Lagrangian 3D tracking of individual bubble swarm members allows a deeper understanding of hydrodynamic bubble-bubble interactions and their collective rise. For multi-view measurements, we have recently developed such a tracking method \citep{hessenkemper20243d}, which is able to track deformable bubbles with low to moderate view obstruction through the bubbles to each other. In the present work, we aim to further enhance the 3D tracking performance by incorporating 2D temporal information in the form of previously established 2D tracks in each view. The new 3D tracking method is able to disambiguate cross-view object associations at each time step by using the 2D track information accumulated over time. In addition, the 2D track information is used in two post-processing steps to improve the completeness of established 3D trajectories. Compared to the previous 3D tracking method, the extended 3D tracking framework shows noticeable improvements in tracking ability, accuracy and completeness of trajectories.