Improved time-frequency distribution using singular value decomposition of Choi-Williams distribution

The Choi-Williams distribution (CWD) is an effective time-frequency (TF) distribution to suppress cross-terms. However, one notable drawback of the CWD is that it fails to suppress cross-terms when either center frequencies of two components are close or their time overlap is significant. In this paper, this problem is addressed by reconstructing TF distribution using basis functions which are extracted through Singular Value Decomposition (SVD) from CWD. The numerical analysis of multi-component Gaussian echoes are presented by using SVD of CWD. It is found that this decomposition and reconstruction approach efficiently eliminate residual cross-terms for which the CWD failed to remove. Results are presented in support of exhibiting the effectiveness of SVD of CWD for cross-term suppression.