Minifying photometric stereo via knowledge distillation-based feature translation.

Photometric stereo (PS) estimates the surface normals of an object by utilizing multiple images captured under different light conditions. To obtain accurate surface normals, a large number of input images is often required. Therefore, a huge effort is needed to capture images and calibrate light directions together with a heavy computational cost. Therefore, in this paper, we propose a robust photometric stereo method even when the number of input images is very small. To this end, we design a feature translation module (FTM) that enriches features having scarce information. In particular, we insert FTMs between the layers of the baseline backbone PS network. Then, activations of each FTM are supervised by distillation loss. For computing distillation loss, we utilize a teacher PS network trained by taking lots of images as inputs. As a result, our PS network requires very few input images but produces a similar quality of output surface normals with the teacher PS network. The proposed method is applicable to both calibrated and uncalibrated PS. We show the effectiveness of the proposed method not only when the number of input images is small but also in various input conditions.

Effects and Mechanism of Surface Water Wettability and Operating Frequency on Response Linearity of Flexible IDE Capacitive Humidity Sensor.

Flexible capacitive humidity sensors are promising for low-cost, wearable, and radio frequency identification sensors, but their nonlinear response is an important issue for practical applications. Herein, the linearity of humidity response was controlled by surface water wettability and operating frequency of sensor, and the mechanism was explained in detail by surface water condensation. For a sensor with a Ag interdigitated electrode (IDE) on a poly(ethylene terephthalate) substrate, the capacitance showed a small linear increase with humidity up to 70% RH but a large nonlinear increase in the higher range. The response linearity was increased by a hydrophobic surface treatment of self-assembled monolayer coating while it was decreased by an ultraviolet/ozone irradiation for hydrophilicity. It was also increased by increasing the frequency in the range of 1-100 kHz, more prominently on a more hydrophilic surface. Based on experiment and simulation, the increase in sensor capacitance was greatly dependent on the geometric pattern (e.g., size, number, and contact angle) and electrical permittivity of surface water droplets. A larger and more nonlinear humidity response resulted from a larger increase in the number of droplets with a smaller contact angle on a sensor surface with higher water wettability and also from a higher permittivity of water at a lower frequency.

Fabrication of truly 3D microfluidic channel using 3D-printed soluble mold.

The field of complex microfluidic channels is rapidly expanding toward channels with variable cross-sections (i.e., beyond simple rounded channels with a constant diameter), as well as channels whose trajectory can be outside of a single plane. This paper introduces the use of three-dimensional (3D) printed soluble wax as cast molds for rapid fabrication of truly arbitrary microfluidic polydimethylsiloxane (PDMS) channels that are not achieved through typical soft lithography. The molds are printed directly from computer-aided design files, followed by simple dissolution using a solvent after molding PDMS, making rapid prototyping of microfluidic devices possible in hours. As part of the fabrication method, the solubility of several build materials in solvents and their effect on PDMS were investigated to remove the 3D-printed molds from inside the replicated PDMS microfluidic channels without damage. Technology limits, including surface roughness and resolution by comparing the designed channels with fabricated cylindrical channels with various diameters, are also characterized. We reproduced a 3D image of an actual human cerebral artery as cerebral artery-shaped PDMS channels with a diameter of 240 µm to prove the developed fabrication technique. It was confirmed that the fabricated vascular channels were free from any leakage by observing the fluorescence fluid fill.

Generation of midfield concentrated beam arrays using periodic metal annular apertures.

Generation of minimally diffracting beam arrays in the midfield region using periodic metal annular apertures is investigated. The relations between the patterns of the diffraction fields and the structural parameters of the periodic metal annular aperture are numerically analyzed. Material dependent transmission characteristics are also studied with finite difference time-domain simulation. The results reveal that the beam concentration efficiency and axial intensity uniformity have a trade-off restriction due to strong inter-aperture interference and surface plasmon mediates the transmission efficiency of the periodic annular apertures. The design criteria of the metal annular aperture to achieve the strong and uniform beam arrays are addressed.