Local Shannon, Rényi, and Tsallis Entropy for Useful Content Extraction from Choi-Williams and Zhao-Atlas-Marks Time-Frequency Distributions

In this paper, we extend the research study on the 2D local entropy method (2DLEM) by applying new entropy measures and time-frequency distributions. In our earlier work, 2D local entropy method is introduced and implemented to the test using both a small sample of real-world data and synthetic signals. We now broaden the scope of our method by applying it to time-frequency distributions that were not introduced in the original paper, as well as changing the entropy measure that is used within the method. Results show that for linear and separate component signals, non-parameterized Shannon entropy is a better choice, while the distributions that were used in the original study outperformed the others.