Function-on-function kriging, with applications to 3D printing of aortic tissues

Abstract

3D-printed medical phantoms, which use synthetic metamaterials to mimic biological tissue, are becoming increasingly important in urgent surgical applications. However, the mimicking of tissue mechanical properties via 3D-printed metamaterial can be difficult and time-consuming, due to the functional nature of both inputs (metamaterial geometry) and outputs (its corresponding mechanical response curve). To deal with this, we propose a novel function-on-function kriging emulation model for efficient tissue-mimicking. For functional inputs, a key novelty of our model is the spectral-distance (SpeD) correlation function, which captures important spectral differences between two functional inputs. Dependencies for functional outputs are then modeled via a co-kriging framework. We further adopt sparse priors on both the input spectra and the output co-kriging covariance matrix, which allows the emulator to learn and incorporate important physics (e.g., dominant input frequencies, output curve properties). Finally, we demonstrate the effectiveness of the proposed SpeD emulator in a real-world study on mimicking human aortic tissue, and show that it can provide quicker and more accurate tissue-mimicking performance compared to existing methods in the medical literature.

A calibration-free method for biosensing in cell manufacturing

“…This variability term δ i (•) is typically unknown. Such a decomposition of a mean trend and a variability term is widely assumed in different modeling methods (see, e.g., Guillas et al, 2018;Chen et al, 2019).…”

Adaptive design for Gaussian process regression under censoring

“…Literature. GP regression (or kriging, see Matheron, 1963) is widely used as a predictive model for expensive experiments (Sacks et al, 1989), and has been applied in many applications, including cosmology (Kaufman et al, 2011), rocket design (Mak et al, 2018), and healthcare (Chen et al, 2019). The key appeals of kriging are its flexible model structure, and its closed-form expressions for prediction and uncertainty quantification.…”