This paper proposes to adapt the Empirical Mode Decomposition Time-Frequency Distribution (EMD-TFD) to non-analytic complex-valued signals. This original method employs the Non uniformly Sampled Bivariate Empirical Mode Decomposition (NSBEMD) to design a filter in the ambiguity domain and clean the Time-Frequency Distribution (TFD) of the signal. This new approach is called NSBEMD-TFD. The suggested adaptation is used in the generation of Inverse Synthetic Aperture Radar (ISAR) image and compared to other Time-Frequency Representation (TFR) such as Spectrogram, Wigner-Ville Distribution (WVD).. .. Furthermore, two criteria to qualify TFD are adjusted to be perform on ISAR images generated by TFD. This method, called NSBEMD-TFD, and those criteria are tested on simulated data and also on data acquired from an anechoic chamber.