ONE REAL-TIME ON-LINE INTELLIGENT WIDTH MEASUREMENT METHOD OF MYOTUBES CULTURED ON SMALL SATELLITES

Spatial environment such as weightlessness may lead to muscle atrophy, which limits the astronauts’ activities in space. To relieve muscle atrophy, some pharmacological experiments about culturing myotubes in vitro will be carried out on small satellites to screen small drug molecules. In the process of differentiation of myotubes, the width distribution of the well-developed myotubes is widely used for observing the influence of small drug molecules. Traditional width measurement methods are measured manually by scientists via softwares, such as Image-Pro Plus, on myotube images. In this paper, we proposed a real-time on-line intelligent method for measuring widths of myotubes. According to the features of the myotube images, high accuracy instance-aware segmentation on myotube images is achieved by adapting one deep learning method. And then, the rough maximum width of each segmented myotube is estimated by distance transformation. Combining the estimated maximum width with the normal line of one edge point on one side of each myotube, the corresponding point on the other side is determined. The width of the myotube at this point is calculated via the Euclidean distance between the two points. We calculated substantial widths at different points of one myotube and choose the maximum one as the maximum width for this myotube. At last, we compared the results of the intelligent width measurement method with the manual method measured by Image-Pro Plus. Experiment results have shown that the proposed method is more accurate.