The accurate identification of the boundaries of the main flow regimes in fluidized beds is very important since the degrees of mixing and mass and heat transfer depend on the prevailing flow regime. Both the minimum fluidization velocity U-mf (0.103 m.s(-1)) and the transition velocity U-trans (0.12 m.s(-1)) to bubbling fluidization regime have been successfully identified based on a newly developed maximum information entropy (IEmax) algorithm applied to gamma photon time series. The latter are recorded by means of gamma-ray densitometry scans performed in an air-polyethylene fluidized bed (0.438 m in ID) operated at ambient conditions. A comparison with the Kolmogorov entropy (KE) algorithm has found that the new approach yielded more accurate transition velocities. The value of U-mf is also validated by means of the profiles of both bed pressure drop and bed height, respectively. It has been found, however, that these parameters are not capable of identifying the second transition velocity U-trans. It is demonstrated that the U-mf value identified on the basis of IEmax is theoretically predictable. In the bubbling fluidization regime, a simple correlation between both the KE and IEmax values is developed.