.png){kind=icon}
.png){kind=icon}
Return
Denoising fluorescent lifetime imaging microscopy images using principal component analysis
.png){kind=icon}
PRE
AI
DOI：10.1117/12.3077519.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Fluorescence lifetime imaging microscopy (FLIM) is a powerful tool that leverages lifetime measurements to enable the study and evaluation of environmental and molecular characteristics of biological samples. As such, there is an increasing demand to develop robust methods to analyze complex FLIM data. Common standard analysis methods such as time domain fitting and fit-free phasor analysis are widely used, however intrinsic noise included in FLIM signals introduces error in the results and adds to the uncertainty in parameters under study. Standard denoising methods such as adaptive and aggressive thresholding and smoothing filtering have limited capability in noise removal and usually work for moderate to high count signals. To circumvent these obstacles, we develop a principal component analysis (PCA)-based method to denoise Poisson-normalized FLIM data efficiently and accurately across different experimental setups. Our approach decomposes the data onto several orthonormal vectors and extracts the underlying signal by removing temporal noise content effectively. We have validated this method using both theoretical and patient-derived colorectal cancer organoid FLIM data and have shown that this method reduces the uncertainty in the data by a factor of up to similar to 5.5. In addition, by avoiding phasor histogram averaging, we reduce signal loss by similar to 50-fold. This approach significantly reduces the standard deviation and improves detection resolution of FLIM data regardless of the imaging system. This exceptional denoising performance and enhanced resolution make it a highly promising tool for exploring biological processes, including metabolic activity, disease dynamics, and treatment response, greatly expanding its relevance in biomedical research.
Keywords:
FLIM
denoising
principal component analysis
lifetime imaging
biomedical research
Journal
M
IF:
0
Papers:
13
Citations:
0