Image-Based Analysis of Patterns Formed in Drying Drops

Image processing and pattern recognition offer a useful and versatile method for optically characterizing drops of a colloidal solution during the drying process and in its final state. This paper exploits image processing techniques applied to cross-polarizing microscopy to probe birefringence and the bright-field microscopy to examine the morphological patterns. The bio-colloidal solution of interest is a mixture of water, liquid crystal (LC) and three different proteins [lysozyme (Lys), myoglobin (Myo), and bovine serum albumin (BSA)], all at a fixed relative concentration. During the drying process, the LC phase separates and becomes optically active detectable through its birefringence. Further, as the protein concentrates, it forms cracks under strain due to the evaporation of water. The mean intensity profile of the drying process is examined using an automated image processing technique that reveals three unique regimes- a steady upsurge, a speedy rise, and an eventual saturation. The high values of standard deviation show the complexity, the roughness, and inhomogeneity of the image surface. A semi-automated image processing technique is proposed to quantify the distance between the consecutive cracks by converting those into high contrast images. The outcome of the image analysis correlates with the initial state of the mixture, the nature of the proteins, and the mechanical response of the final patterns. The paper reveals new insights on the self-assembly of the macromolecules during the drying mechanism of any aqueous solution.