Xiang Zhang

Bio: Xiang Zhang is an academic researcher from Baylor College of Medicine. The author has contributed to research in topics: Medicine & Computer science. The author has an hindex of 154, co-authored 1733 publications receiving 117576 citations. Previous affiliations of Xiang Zhang include University of California, Berkeley & University of Texas MD Anderson Cancer Center.

Heat and mass transfer in pulsed-laser-induced phase transformations

Abstract: Publisher Summary This chapter reviews a recent research on pulsed-laser-induced phase-change transformations at the nanosecond time scale. The melting of semiconductor materials was probed by optical reflectance, transmittance, electrical conductance, and infrared pyrometry. The experimental results were in general agreement with the thermal model, except for the high-irradiance regime. Direct measurements of the solid-liquid interface in melting of polysilicon films on glass substrates showed superheat by over 100° C. The rapid mass transfer in the liquid silicon phase allowed successful fabrication of well-controlled and sharply defined box-shaped ultrashallow dopant junction profiles. The chapter discusses that utilizing heat transfer and stability analysis, the generation of surface morphology was attributed to Rayleigh-Taylor instability triggered by the reduction of density on melting. Time-of-flight measurements of the kinetic-energy distribution in near-threshold sputtering of gold produced values exceeding the thermal model expectations. A computational heat-transfer analysis of the heat transfer and the vapor gas dynamics adopting the transparent vapor assumption are also discussed.

Brownian motion of tethered nanowires.

Abstract: Brownian motion of slender particles near a boundary is ubiquitous in biological systems and in nanomaterial assembly, but the complex hydrodynamic interaction in those systems is still poorly understood. Here, we report experimental and computational studies of the Brownian motion of silicon nanowires tethered on a substrate. An optical interference method enabled direct observation of microscopic rotations of the slender bodies in three dimensions with high angular and temporal resolutions. This quantitative observation revealed anisotropic and angle-dependent hydrodynamic wall effects: rotational diffusivity in inclined and azimuth directions follows different power laws as a function of the length, \ensuremath{\sim}${L}^{\ensuremath{-}2.5}$ and \ensuremath{\sim}${L}^{\ensuremath{-}3}$, respectively, and is more hindered for smaller inclined angles. In parallel, we developed an implicit simulation technique that takes the complex wire-wall hydrodynamic interactions into account efficiently, the result of which agreed well with the experimentally observed angle-dependent diffusion. The demonstrated techniques provide a platform for studying the microrheology of soft condensed matters, such as colloidal and biological systems near interfaces, and exploring the optimal self-assembly conditions of nanostructures.

Growth of high-quality GaAs/AlAs Bragg mirrors on patterned InP-based quantum well mesa structures

Abstract: We have investigated the regrowth of GaAs/AlAs quarter-wave Bragg reflectors on patterned mesa InP-based quantum well heterostructures that can be fabricated into 1.55 μm vertical cavity surface emitting lasers. It is seen from transmission electron and scanning electron microscopy that the multiple layer GaAs-based mirrors can be grown on InP-based heterostructure mesas of diameters 10–40 μm without noticeable propagation of defects into the reflector layers or the quantum well region below. At the same time the photoluminescence from the quantum wells after regrowth indicates that lasers can be fabricated.

In Vivo Imaging of Microvasculature during Anesthesia with High-Resolution Photoacoustic Microscopy.

Abstract: Anesthesia monitoring is extremely important in improving the quality of anesthesia and ensuring the safety of patients in operation. Photoacoustic microscopy (PAM) is proposed to in vivo image the skin microvasculature of 10 nude mice undergoing general anesthesia by using the isoflurane gas with a concentration of 3%. Benefiting from strong optical absorption of hemoglobin, PAM has good contrast and high resolution in mapping of microvasculature. A series of high quality images can clearly reveal the subtle changes of capillaries in morphology over time. Two indices, vessel intensity and vessel density, are extracted from these images to measure the microvasculature quantitatively. The imaging results show that the vessel intensity and density are increased over time. After 65 min, the vessel intensity increased 42.7 ± 8.6% and the density increased 28.6 ± 12.2%. These indices extracted from photoacoustic images accurately reflect the greater blood perfusion undergoing general anesthesia. Additionally, abnormal reductions of vessel intensity and density are also observed as overtime anesthesia. This preclinical study suggests that PAM holds potential to monitor anesthesia by imaging the skin microvasculature.

Negative group velocity of surface plasmons on thin metallic films

Abstract: By tailoring the dispersion curve of surface plasmons (SPs) of a thin metallic film surrounded by dielectric half-spaces, it is shown that the group velocity of the symmetric mode is always positive, while the group velocity of the anti-symmetric mode can be negative. Consequently, the forward and backward propagation of SPs, in which the energy flow is respectively parallel or antiparallel to the wave vector, can be realized. The physical origin of the intriguing backward SPs is given. Furthermore, schemes for the negative refraction and imaging of SPs are proposed by incorporating two plasmon modes with group velocities of opposite signs.

Deep Residual Learning for Image Recognition

Abstract: Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers—8× deeper than VGG nets [40] but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions1, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.

Very Deep Convolutional Networks for Large-Scale Image Recognition

Abstract: In this work we investigate the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting. Our main contribution is a thorough evaluation of networks of increasing depth using an architecture with very small (3x3) convolution filters, which shows that a significant improvement on the prior-art configurations can be achieved by pushing the depth to 16-19 weight layers. These findings were the basis of our ImageNet Challenge 2014 submission, where our team secured the first and the second places in the localisation and classification tracks respectively. We also show that our representations generalise well to other datasets, where they achieve state-of-the-art results. We have made our two best-performing ConvNet models publicly available to facilitate further research on the use of deep visual representations in computer vision.

Deep Residual Learning for Image Recognition

Abstract: Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers---8x deeper than VGG nets but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.

Going deeper with convolutions

Abstract: We propose a deep convolutional neural network architecture codenamed Inception that achieves the new state of the art for classification and detection in the ImageNet Large-Scale Visual Recognition Challenge 2014 (ILSVRC14). The main hallmark of this architecture is the improved utilization of the computing resources inside the network. By a carefully crafted design, we increased the depth and width of the network while keeping the computational budget constant. To optimize quality, the architectural decisions were based on the Hebbian principle and the intuition of multi-scale processing. One particular incarnation used in our submission for ILSVRC14 is called GoogLeNet, a 22 layers deep network, the quality of which is assessed in the context of classification and detection.