Abstract: A method of training AI for label-free cell viability determination includes a step of providing a cell sample, a step of obtaining a fluorescence image and a DHM image of the cell sample, a step of determining a first cell viability of the cell sample according to the fluorescence image of the cell sample, a step of labeling the DHM image of the cell sample as a model specifying the first cell viability, and a step of performing AI training by using the model containing the DHM image of the cell sample.

Abstract: A device and method for detecting crystal quality of a low temperature poly-silicon (LTPS) backplane are provided, the method including: projecting narrowband light to the LTPS backplane; performing image capturing of each position on a surface of the LTPS backplane at a first angle in a first axial direction to obtain a first diffraction image, the first angle being an angle of maximum diffraction light intensity in the first axial direction; performing another image capturing of each position on the surface of the LTPS backplane at a second angle in a second axial direction to obtain a second diffraction image, the second angle being an angle of maximum diffraction light intensity in the second axial direction; and determining the crystal quality of the LTPS backplane based on a diffraction light intensity distribution obtained from the first and second diffraction images.