Blood vessel detection using skin impedance tomography and spectroscopy.

Blood vessel detection is an essential method in a non-invasive circulatory monitoring system. However, only using camera-based image processing for finding a vein has many limitations due to the complexity of underneath tissue. Impedance tomography and spectroscopy can collect electrical characteristics map of the biological tissue precisely. The study developed instrumentation to measure the impedance with eight small electrodes, which can observe near vein area and show the conductivity map with small patches using tomographic reconstruction. The study examined the cephalic vein of a healthy human arm to confirm the feasibility of vein location recognition. The system could perform 86.8% pixel accuracy and achieve a mIoU score of 63.7% for vessel location segmentation in 5cm2 area observation. The system could identify a human blood vessel's electrical characteristic and visualize the passage.Clinical Relevance- This approach will support intravascular therapy by identifying the vessel location automatically and providing information to the vessel monitoring system in backend. In addition, the system would allow efficient data entry into the electronic medical record for management of hospitalized patients.