| Abstract: | A frequency‐domain (FD) analysis of fluorescence lifetime (FLT) is a unique and rapid method for cellular and intracellular classifications that can serve for medical diagnostics purposes. Nevertheless, its data analysis process demands nonlinear fitting algorithms that may distort the resolution of the FLT data and hence diminish the classification ability of the method. This research suggests a sample classification technique that is unaffected by the analysis process as it is based on the squared distance (D2) between the raw frequency response data (FRD). In addition, it presents the theory behind this technique and its validation in two simulated data sets of six groups with similar widely and closely spaced FLT data as well as in experimental data of 43 samples from bacterial and viral infected and non‐infected patients. In the two simulated tests, the classification accuracy was above 95% for all six groups. In the experimental data, the classification of 41 out of 43 samples matched earlier report and 29 out of 31 agreed with preliminary physician diagnosis. The D2 approach has the potential to promote FD‐time resolved fluorescence measurements as a medical diagnostic technique with high specifity and high sensitivity for many of today's conventional diagnostic procedures.  |
