Peter is correct.

Additionally, the two components in FLIM-FRET may be considered as a mixture of donor-only (highest lifetime in your image) and single (or close to single) FRET efficiency(lower lifetime component). However this information is given for the donor fluorescence decay over the integration of the image. Fluorescence decays of single pixels in your image may be too noisy.

For this you can try using Enrico Gratton's Phasor approach to identify one of 3:

1. Pixels that contain no FRET - points on the 'universal semi-circle'

2. Pixels that contain 100% of molecules overgoing FRET - points on the 'universal semi-circle'

3. Pixels that contain a mixture of both 1 & 2  - points inside the 'universal semi-circle' on a line connecting between the two points from 1 & 2

Why do you want to do that? There are other simpler methods of obtaining that. One need not do FILM-FRET for that, as rightly pointed out by Peter.

Alternatively you could use Gratton's Phasor approach as pointed by Eitan, but requires more effort. Ultimately it depends on your objective.

Thank you so much all for the valuable suggestions. I am considering other methods too, including Sensitized Emission (SE).  I was wondering if we can use FLIM-FRET in other cases. Thank you again.

For acceptor sensitized emission, I really hope you have negligible contributions in the acceptor detection channel from donor fluorescence leakage and from direct acceptor excitation.

Why not measure donor FLIM, then intentionally bleach your acceptor, then measure donor lifetime again (and hopefully the sample did not change position drastically and you are left with a donor-only sample