Characterization of Bioengineered Tissues by Digital Holographic Vibrometry and 3D Shape Deep Learning

One of the critical components of large-scale manufacturing of bioengineered tissues is the sensing of information for quality control and critical feedback of tissue growth. Modern sensors that measure mechanical qualities of tissues, however, are invasive and destructive. The goal of this project is to develop noninvasive methodologies to measure the mechanical properties of tissue engineering products. Our approach is to utilize acoustic waves to induce nanoscale level vibrations in the engineered tissues in which corresponding displacements are measured in full-field with quantitative optical techniques. A digital holographic system images the tissue’s vibration at significant modes and provides the displacement patterns of the tissue. These data are used to train a supervised learning classifier with a goal of using the comparisons between the experimental vibrational modes and the ones obtained by finite element simulation to estimate the physical properties of the tissue. This methodology has the promise of mechanical properties that would allow technicians to noninvasively determine when samples are ready to be packaged, if their growth deviates from expected time frames, or if there are defects in the tissue. It is expected that this approach will streamline several components of the quality control and production process.