The ELF or fluorescence-labeled enzyme activity (FLEA) technique is a culture-independent single-cell tool for assessing plankton enzyme activity in close-to-in situ conditions. We demonstrate that single-cell FLEA quantifications based on two-dimensional (2D) image analysis were biased by up to one order of magnitude relative to deconvolved 3D. This was basically attributed to out-of-focus light, and partially to object size. Nevertheless, if sufficient cells were measured (25-40 cells), biases in individual 2D cell measurements were partially compensated, providing useful and comparable results to deconvolved 3D. We also discuss how much caution should be used when comparing the single-cell enzyme activities of different sized bacterio- and/or phytoplankton populations measured on 2D images. Finally, a novel method based on deconvolved 3D images (wide field restoration microscopy; WFR) was devised to improve the discrimination of similar single-cell enzyme activities, the comparison of enzyme activities between different size cells, the measurement of low fluorescence intensities, the quantification of less numerous species, and the combination of the FLEA technique with other single-cell methods. These improvements in cell enzyme activity measurements will provide a more precise picture of individual species' behavior in nature, which is essential to understand their functional role and evolutionary history.