arXiv subjects

Chih-Fan Chen

Publications and source records attributed to Chih-Fan Chen.

4 recordsLinked to original sources

arXiv2024

Spatially Resolved Kinematics of SLACS Lens Galaxies. I: Data and Kinematic Classification

We obtain spatially resolved kinematics with the Keck Cosmic Web Imager (KCWI) integral-field spectrograph for a sample of 14 massive (11 < log$_{10}$ M$_*$/M$_{\odot}$ < 12) lensing early-type galaxies at z~0.15-0.35 from the Sloan Lens ACS (SLACS) Survey. We integrate kinematic maps within the effective radius and examine rotational and dispersion velocities, showing that 11/14 are slow rotators. The dataset is unprecedented for galaxy-scale strong lenses in terms of signal-to-noise ratio (S/N), sampling, and calibration. Systematics are at 1-1.4%, and positive covariance is <1% between sample galaxies and between spatial bins, with primary contibutions from stellar template library selection and fitted wavelength range. This enables cosmographic inference with lensing time delays with <2% uncertainty on H$_0$. We integrate the datacubes within various circular apertures and compare with SDSS velocity dispersions. Velocity dispersions extracted from SDSS spectra for these 14 galaxies, which have low S/N (~9/$\r{A}$) relative to the parent sample, are subject to systematic errors (and covariance) due to stellar template library selection at the level of 3(2)%, which need to be added to the random errors. Comparison between our KCWI measurements, our analysis of SDSS spectra, and previously published measurements based on SDSS spectra shows mean differences within a few percent, which are insignificant given the uncertainties of the SDSS-based measurements. Correlations between scaling relations using quantities inferred from dynamical, lensing, and stellar population models agree with previous SLACS analysis with no statistically significant change. A follow-up paper will present Jeans modeling in the context of broader studies of galaxy evolution and cosmology.

astro-ph.CO↗

arXiv2019

Face Beautification: Beyond Makeup Transfer

Facial appearance plays an important role in our social lives. Subjective perception of women's beauty depends on various face-related (e.g., skin, shape, hair) and environmental (e.g., makeup, lighting, angle) factors. Similar to cosmetic surgery in the physical world, virtual face beautification is an emerging field with many open issues to be addressed. Inspired by the latest advances in style-based synthesis and face beauty prediction, we propose a novel framework of face beautification. For a given reference face with a high beauty score, our GAN-based architecture is capable of translating an inquiry face into a sequence of beautified face images with referenced beauty style and targeted beauty score values. To achieve this objective, we propose to integrate both style-based beauty representation (extracted from the reference face) and beauty score prediction (trained on SCUT-FBP database) into the process of beautification. Unlike makeup transfer, our approach targets at many-to-many (instead of one-to-one) translation where multiple outputs can be defined by either different references or varying beauty scores. Extensive experimental results are reported to demonstrate the effectiveness and flexibility of the proposed face beautification framework.

cs.CV↗

arXiv2019

Digital Twin: Acquiring High-Fidelity 3D Avatar from a Single Image

We present an approach to generate high fidelity 3D face avatar with a high-resolution UV texture map from a single image. To estimate the face geometry, we use a deep neural network to directly predict vertex coordinates of the 3D face model from the given image. The 3D face geometry is further refined by a non-rigid deformation process to more accurately capture facial landmarks before texture projection. A key novelty of our approach is to train the shape regression network on facial images synthetically generated using a high-quality rendering engine. Moreover, our shape estimator fully leverages the discriminative power of deep facial identity features learned from millions of facial images. We have conducted extensive experiments to demonstrate the superiority of our optimized 2D-to-3D rendering approach, especially its excellent generalization property on real-world selfie images. Our proposed system of rendering 3D avatars from 2D images has a wide range of applications from virtual/augmented reality (VR/AR) and telepsychiatry to human-computer interaction and social networks.

cs.CV↗

arXiv2018

Solving the paradox of the folded falling chain by considering horizontal kinetic energy and link geometry

A folded chain, with one end fixed at the ceiling and the other end released from the same elevation, is commonly modeled as an energy-conserving system in one-dimension. However, the analytical paradigms in previous literature is unsatisfying: The theoretical prediction of the tension at the fixed end becomes infinitely large when the free end reaches the bottom, contradicting to the experimental observations. Furthermore, the dependence of the total falling time on the link number demonstrated in numerical simulations is still unexplained. Here, considering the horizontal kinetic energy and the geometry of each link, we derived analytical solutions of the maximal tension as well as the total falling time, in agreement with simulation results and experimental data reported in previous studies. This theoretical perspective shows a simple representation of the complicated two-dimensional falling chain system and, in particular, specifies the signature of the chain properties.

physics.class-ph↗