Deep learning enabled in vitro predicting biological tissue thickness using force measurement device.

Area of Science:

• Biomedical Engineering
• Materials Science
• Artificial Intelligence

Background:

• Accurate biological tissue (BT) thickness measurement is crucial for medical diagnosis and animal nutrition.
• Traditional methods are complex, expensive, and induce biological stress.

Purpose of the Study:

• To develop a novel, non-invasive in vitro method for measuring biological tissue thickness.
• To integrate a force test system (FST) with a deep learning model for enhanced accuracy.

Main Methods:

• Proposed an in vitro approach combining a force test system (FST) and a discrete multiwavelet transform convolutional neural network (DMWA-CNN) prediction model.
• Conducted comprehensive experiments and model comparisons to validate the approach.
• Tested the method's robustness against variations in elastic modulus, external load, and small thickness differences.

Main Results:

• The DMWA-CNN model achieved 100% accuracy for artificial biological tissues, outperforming traditional algorithms.
• The proposed approach demonstrated robustness to variations in elastic modulus (E), external load (F), and small thickness differences (Ts).
• Experimental measurement of four pork tissue types yielded accuracy not less than 98.2%.

Conclusions:

• The FST integrated with the DMWA-CNN algorithm offers a highly accurate and robust method for in vitro biological tissue thickness measurement.
• This novel approach shows potential for application in biomechanical parameter prediction.
• The developed method overcomes limitations of traditional techniques, offering a cost-effective and less invasive alternative.