A Survey on Long-Tailed Visual Recognition

The heavy reliance on data is one of the major reasons that currently limit the development of deep learning. Data quality directly dominates the effect of deep learning models, and the long-tailed distribution is one of the factors affecting data quality. The long-tailed phenomenon is prevalent due to the prevalence of power law in nature. In this case, the performance of deep learning models is often dominated by the head classes while the learning of the tail classes is severely underdeveloped. In order to learn adequately for all classes, many researchers have studied and preliminarily addressed the long-tailed problem. In this survey, we focus on the problems caused by long-tailed data distribution, sort out the representative long-tailed visual recognition datasets and summarize some mainstream long-tailed studies. Specifically, we summarize these studies into ten categories from the perspective of representation learning, and outline the highlights and limitations of each category. Besides, we have studied four quantitative metrics for evaluating the imbalance, and suggest using the Gini coefficient to evaluate the long-tailedness of a dataset. Based on the Gini coefficient, we quantitatively study 20 widely-used and large-scale visual datasets proposed in the last decade, and find that the long-tailed phenomenon is widespread and has not been fully studied. Finally, we provide several future directions for the development of long-tailed learning to provide more ideas for readers.

     

Absorption effect in the calculation of a second-order nonlinear coefficient from the data of a maker fringe experiment

The effect of light absorption by sample in the analysis of Makerfringe data for estimating a second-order nonlinear coefficient hasbeen studied experimentally. Two theories, one by Jerphagnon andKurtz that neglects the absorption effect and one by Herman and Haydenthat takes into account the absorption effect, were compared with theexperimental results. It was found that Jerphagnon and Kurtz'sformula was unable to predict correctly not only the magnitude but alsothe incident angle dependence or the sample thickness dependence of thesecond harmonic signal generated by the sample with strong absorption, whereas the theory by Herman and Hayden was able to make thosepredictions fairly well. It was also found that the error in theestimated nonlinear coefficient when one uses Jerphagnon and Kurtz'sformula could be as large as 2-4 times the true value, depending onsample thickness.     

A study of a new data association and track initiation method with normalized distance squared ordering

In an underwater environment, measurements regarding true targets and false targets (clutter) can be made. Therefore, a suitable data association method to exactly detect and track a target and an efficient track initiation method for judging tracks formed by the target should be selected in this environment. This paper attempts to propose a new data association method and track initiation method to detect and track targets more effectively. Also, the performance of the new method is tested in a series of Monte Carlo simulation runs and is compared with the existing data association and track initiation methods in a cluttered environment.     

基于SOPC的RSA密码芯片设计

为了提高RSA密码算法在密码芯片上的实现性能,基于SOPC技术设计了一款RSA密码芯片,完成了NIOS Ⅱ的软硬件协同设计.该硬件系统由处理器、存储模块和自定义指令3部分组成,其关键的模乘模块采用改进的模乘算法和交叉存取设计方案设计实现,显著提高了系统性能.整体设计在ALTERA的CYCLONEⅡ器件上进行了验证与测试,并进行了逻辑综合及布局布线.与传统的RSA密码芯片设计相比,该芯片系统灵活性高,资源占用少.     

Dynamically modulated intensity interrogation scheme using waveguide coupled surface plasmon resonance sensors

An electro-optically modulated intensity interrogation method based on tunable waveguide coupled surface plasmon resonance sensors has been proposed. It has been theoretically and experimentally demonstrated that the proposed scheme can enable sensitive measurement of measurand variations. By modulating the refractive index in the waveguide layer, this interrogation method yields modulated signal whose amplitude is related to measurand's refractive index. This amplitude modulated signal offers a higher signal to noise ratio and eliminates additive noise in the sensor system. A preliminary investigation using saline buffers with different NaCl concentrations shows a resolution of 2.3 × 10^?6 refractive index unit by our approach. Resolution can be controlled by the amplitude of the applied modulation voltage and can be further enhanced by optimizing the device structure or improving the electro-optical (E-O) coefficient of the E-O material. This approach is simple, stable, and promising for low-cost or multi-channel SPR biosensor applications.     

Synthesis and characterization of novel blue azo‐dye derivatives containing acrylate group for dye‐based color filters

Abstract— Even though dyes have a fine resolution and good chromaticities, they are not widely used as coloring materials for color filters (CFs) due to their low thermal stability and chemical resistance. A series of azo‐dye derivatives, which consist of two cross‐linkable acrylate or methacrylate groups to improve thermal and chemical properties, have been synthesized and used to fabricate color filters. The spectral properties and chemical/thermal stabilities of the fabricated CFs were investigated by comparing dye‐based CFs, without a complicated dispersion process, but with pigment‐based CFs using dispersed pigment. Also, more properties including the development test and surface morphologies lithographic properties were studied. The synthesized azo dyes were characterized by elemental analysis, UV‐visible spectra, IR, mass, and ¹H‐NMR spectra.     

Efficient aerodynamic shape optimization of transonic wings using a parallel infilling strategy and surrogate models

Surrogate models are used to dramatically improve the design efficiency of numerical aerodynamic shape optimization, where high-fidelity, expensive computational fluid dynamics (CFD) is often employed. Traditionally, in adaptation, only one single sample point is chosen to update the surrogate model during each updating cycle, after the initial surrogate model is built. To enable the selection of multiple new samples at each updating cycle, a few parallel infilling strategies have been developed in recent years, in order to reduce the optimization wall clock time. In this article, an alternative parallel infilling strategy for surrogate-based constrained optimization is presented and demonstrated by the aerodynamic shape optimization of transonic wings. Different from existing methods in which multiple sample points are chosen by a single infill criterion, this article uses a combination of multiple infill criteria, with each criterion choosing a different sample point. Constrained drag minimizations of the ONERA-M6 and DLR-F4 wings are exercised to demonstrate the proposed method, including low-dimensional (6 design variables) and higher-dimensional problems (up to 48 design variables). The results show that, for surrogate-based optimization of transonic wings, the proposed method is more effective than the existing parallel infilling strategies, when the number of initial sample points are in the range from N _v to 8N _v (N _v here denotes the number of design variables). Each case is repeated 50 times to eliminate the effect of randomness in our results.