Coding Optimization for Fast Fluorescence Lifetime Imaging

Fluorescence lifetime imaging (FLIM) is used for measuring material properties in a wide range of applications, including biology, medical imaging, chemistry, and material science. In frequency-domain FLIM (FD-FLIM), the object of interest is illuminated with a temporally modulated light source. The fluorescence lifetime is measured by computing the correlations of the emitted light with a demodulation function at the sensor. The signal-to-noise ratio (SNR) and the acquisition time of a FD-FLIM system is determined by the coding scheme (modulation and demodulation functions). We develop theory and algorithms for designing high-performance FD-FLIM coding schemes that can achieve high SNR and short acquisition time, given a fixed source power budget. Based on a geometric analysis of the image formation and noise model, we propose a novel surrogate objective for the performance of a given coding scheme. The surrogate objective is extremely fast to compute, and can be used to efficiently explore the entire space of coding schemes. Based on this objective, we design novel, high-performance coding schemes that achieve up to an order of magnitude shorter acquisition time as compared to existing approaches. We demonstrate the performance advantage of the proposed schemes in a variety of imaging conditions, using a modular hardware prototype that can implement various coding schemes.


Coding Optimization for Fast Fluorescence Lifetime Imaging

Jongho Lee, JV Chacko, B Dai, SA Reza, AK Sagar, KW Eliceiri, A Velten, Mohit Gupta

ACM TOG, presented at SIGGRAPH 2019

Space of captured intensities for FD-FLIM and C-ToF imaging

(a) In FD-FLIM, the set of intensity points when the lifetime is varied forms an open 1D curve, with a non-uniform inter-point distance. (b) In C-ToF imaging, the set of intensity points when the scene depth is varied form a closed loop with a uniform inter-point distance

Surrogate of mean lifetime error

The surrogate values have a strong correlation with the mean lifetime errors across various coding schemes and frequencies, and thus, can be used as a light-weight alternative to the computationally intensive mean lifetime error, in order to find the optimal coding scheme efficiently.

Visual comparisons of fluorescence lifetime images

(a) A fluorescent sample with two different fluorescence lifetimes for the foreground and the background is excited by a low power light source. (b) With the conventional coding scheme and 0.8 ms/pixel acquisition time, no clear boundary is observed between the foreground and background (c) A considerably longer (10 ms/pixel) acquisition time is required to obtain a clear boundary. (d) With the proposed coding schemes, 0.8 ms/pixel acquisition time is sufficient to detect a clear boundary.

Data acquisition time and lifetime precision comparison

Relationship between the mean absolute error, MAE of the measured lifetimes for Rhodamine 110 in ethanol with data acquisition time. Expo - Square requires considerably shorter acquisition time than Sinusoid - Sinusoid to achieve the same SNR.


Presentation Slides


Share This Article

Share on LinkedInShare on FacebookTweet about this on Twitter