The Beckman Institute Graduate Student Seminar Series presents the work of outstanding graduate students working in Beckman research groups. The seminar starts at Noon in Beckman Institute Room 5602 and is open to the public. Lunch will be served.
We proposed an analysis-by-synthesis framework for 3D face reconstruction and face recognition with variant pose, illumination and expression (PIE). First, an efficient 2D-to-3D integrated face reconstruction approach is introduced to reconstruct a personalized 3D face model from a single frontal face image with neutral expression; Then, realistic virtual faces with different PIE are synthesized based on the personalized 3D face to characterize the face subspace; Finally, face recognition is conducted based on these representative virtual faces. Compared with other related works, this framework has the following advantages: 1) only one single frontal face is required for face recognition, which avoids the burdensome enrollment work; 2) the synthesized face samples provide the capability to conduct recognition under difficult conditions like complex PIE; and 3) the proposed 2D-to-3D integrated face reconstruction approach is fully automatic and more efficient, which can also be driven for facial animation and emotive speech avatar. The extensive experimental results show that the synthesized virtual faces significantly improve the accuracy of face recognition with variant PIE. Quantitative evaluations are also conducted on the reconstruction result to show its strength and limitation for further improvement.
Linear Response Imaging of the dynamical hydration structure surrounding aqueous electrons
Water-mediated interactions drive a number of chemical, physical and biological processes in solution. Examples of this span from aqueous chemical reactions to self-assembly of biomolecular structures, all of which are affected by local density fluctuations in the surrounding hydration structure. We have developed a novel technique to image the dynamical density fluctuations in liquid water around atomic-scale charge densities. We combine the ability of inelastic x-ray scattering measurements to map out the spatio-temporal dynamics of bulk water with linear response theory to calculate the hydration structure around arbitrary dynamical charge densities. Further, we use this technique to investigate the dynamics of water around a hydrated electron, which to this point has only been simulated. Future applications of this technique will also be discussed.
Nanometer-scale Probes for Scanned Probe Microscopy Produced by Field-Directed Sputter Sharpening
Ultra-sharp needle-like probes are of interest for electron emission and atomic-resolution imaging and patterning in the scanning tunneling microscope (STM). In our work, the use of such ultra-sharp probes improves the imaging capabilities of the STM and the ability to pattern nanostructures and molecular devices with atomic-precision. In meeting this need, we have developed a Field-Directed Sputter Sharpening (FDSS) technique for producing sharp metallic probes on the nanometer and atomic scale. The FDSS technique produces probes that are extremely sharp when compared to classical sputter erosion results, is compatible with massively parallel batch processing and regeneration of damaged probes, is self-limiting in nature, and is compatible with a wide range of materials. We demonstrate our ability to sharpen tungsten, platinum-iridium alloy, and hafnium diboride, a novel metallic-ceramic material of extreme hardness. We find that these probes provide stable imaging and enable extremely precise atomic patterning of the hydrogen-passivated silicon surface.