Fig. 1: Illustration of full-field displacement estimation. | npj Artificial Intelligence

Fig. 1: Illustration of full-field displacement estimation.

From: Estimating full-field displacement in biological images using deep learning

Fig. 1

(a) Reference (Ref.) and (b) deformed (Def.) fluorescence microscopy images of mouse cardiomyocytes. The full-field displacement from the reference to the deformed state can be represented by a (c) vector image or (d) phasor plot, where the hue represents the angle and saturation represents the magnitude of the displacement vector. Scale bars are 10 μm. e Visual demonstration of the mapping from displacement to colour. (f) Graphical representation of the (i, ii) deep learning method and (iii, iv) loss functions, where u is the displacement vector field and the terms Real and Fake refer, respectively, to real microscopy data and simulated data. (g) Comparison of the performance displacement estimation methods in real data quantified using the peak signal-to-noise ratio (PSNR).

Back to article page