We invented a non-destructive terahertz (THz) deep-learning computed tomography system based on time-domain spectroscopy. In the method, THz time-domain signals are profiled. Multiple features are retrieved from those profiles by a trained model and transformed to the spatial domain to reconstruct a cross-sectional tomographic image. We have also invented a 3D THz tomographic system based on multi-scale spatio-spectral feature fusion in a multi-scale manner. We believe our work will stimulate further applicable research of THz tomographic imaging with advanced computer vision techniques.

We invented a non-destructive terahertz deep-learning tomography system based on plasmonic THz photoconductive antennas. Under a low optical pump power of 1.4mW, we experimentally demonstrate 105W broadband terahertz radiation in 0.1-2 THz frequency range. To improve the spatial and contrast resolution, we invented time domain spectroscopy deep learning model to achieve near 100％ spatial and frequency data utility rate and better spatial resolution. Compared to the conventional algorithm based on spatial information, the model has achieved 46.7％ better in the mean square error index.

With the advancement of medical technology, countries are gradually entering an aging society. As a result, the demand for medical care for aging and chronic diseases is increasing progressively, driving the growth of the global medical equipment market. Terahertz technology is very suitable for the high-end imaging medical material industry because of its non-invasive, non-destructive, non-ionizing, material-classification. By combining the terahertz technology with industries, we can upgrade traditional industries, train outstanding talents, and attract foreign capital.

線上展網址：
https://tievirtual.twtm.com.tw/iframe/efec2bcf-09f0-4efe-9b3b-3d83d6c71381?group=23bfb1fa-dd5b-4836-81a1-4a1809b1bae5&lang=en