Fluorescence Excitation-Emission-Matrix Imaging .
Oren KatzTravis FergusonEmma AbbeySarah-Johanna KloseChris PrüfertHans-Peter LoockPublished in: Analytical chemistry (2023)
We present a 4-dimensional (4D) fluorescence imaging system in which each of the 65,536 pixels in the image array contains an excitation-emission-matrix spectrum with 31 excitation wavelengths and 8 emission wavelengths ( x , y , λ exc , λ em ). Hadamard-transform multiplexing of the excitation light from a 31-channel programmable light source allows for an increase in the data acquisition rate so that each 65,536-pixel image can be obtained within 8 s. The system is demonstrated and characterized using, first, a 4D image of 10 capillaries filled with four dye solutions and their binary and ternary mixtures, and, second, using a sequence of about 100 images of layered fluorescent dye solutions and their changing fluorescence as a function of temperature. Multivariate analysis using parallel factor analysis produces images of the spatial distribution of the fluorophores together with their relative intensity as a function of time.
Keyphrases
- energy transfer
- deep learning
- fluorescence imaging
- quantum dots
- convolutional neural network
- high resolution
- highly efficient
- artificial intelligence
- photodynamic therapy
- ionic liquid
- single molecule
- machine learning
- optical coherence tomography
- reduced graphene oxide
- data analysis
- high throughput
- high intensity
- electronic health record