[A method for measuring the thickness of biological tissues based on piezoelectric effect and deep learning]

Given the issues of traditional methods in measuring biological tissue thickness, including complexity, high cost and vulnerability to trigger biological stress responses, this paper proposes a piezoelectric method for measuring the thickness of in vitro biological tissue based on discrete multiwavelet Transform combined with convolutional neural network (DMWT-CNN). In view of this, this article conducted multiple comprehensive experiments and model comparisons to demonstrate the superiority of the proposed method. The model comparison results show that the DMWT-CNN model proposed in this paper has higher accuracy in thickness estimation compared with other deep learning algorithms, including Morlet-Inception hybrid neural network and fast Fourier transform-convolutional neural network (FFT-CNN). At the same time, it achieved 100% accuracy in measuring the thickness of artificial biological tissues. In addition, the experimental results indicate that the piezoelectric in vitro biological tissue thickness measurement method based on DMWT-CNN is robust to changes in elastic modulus, external load changes, and small thickness differences. Finally, this article conducted experimental measurements on the thickness of ex vivo pork with four different thicknesses, and the results showed that their accuracy values were not less than 98.2%, providing a novel technical approach for the precise measurement of physical property parameters in biological tissues.