Fig. 6From: Intelligent fluorescence image analysis of giant unilamellar vesicles using convolutional neural networkContour intensity analysis for the determination of phase separation state. a A typical example of GUV detection from a uniform vesicle sample. Blue circles indicate the detected GUVs classified as uniform or no-phase domains. The composition of the vesicle was DOPC:Ni-DGS = 90:10 with 0.2 mol% of TopFluor-Cholesterol replacing the DOPC. b A typical example of GUV detection from a phase-separated vesicle sample. Yellow circles indicate GUVs classified as phase separated or the coexistence of phase domains. The composition of the vesicle was DOPC:Ni-DGS:DPPC:Cholesterol = 15:10:50:25 with 0.2 mol% of TopFluor-Cholesterol replacing the cholesterol. The images shown in a and b are 127.3 µm in width and height. c An example intensity trace along the perimeter of a phase-separated vesicle. The intensity trace shows clear discontinuity at the boundaries of two-phase domains, which can be used to determine the existence of phase domains. The average intensity in each domain can also be estimated from the intensity analysis, as shown. d An example intensity trace along the perimeter of a uniform vesicle. The intensity trace shows no detectable discontinuity, which suggests the nonexistence of phase domains within the vesicleBack to article page