基于U-Net的盐体识别方法

卷积神经网络在计算机视觉领域取得重大突破，利用其强大的图像处理能力，将地下沉积盐体的识别问题转化为图像语义分割问题，应用深度卷积神经网络实现盐体地震图像的像素级语义分割.本文在U-Net基础上，增加网络深度并同时引入批归一化和Dropout处理，使得神经网络模型具有更高的可信度和更强的泛化能力.通过实验发现，在卷积层之后引入批归一化处理，并在池化层和叠加层之后引入Dropout可以稳定提升模型对盐体图像的分割性能.     

靶丸支撑膜对ICF内爆性能影响的模拟研究

使用二维多群辐射扩散流体力学程序LARED-S,模拟研究靶丸支撑膜在惯性约束聚变氘氘（DD）气体靶内爆过程中的扰动演化过程及其对内爆性能的影响.二维模拟表明：靶丸支撑膜显著降低DD气体靶内爆的中子产额,二维模拟产额为一维结果的55.2%.内爆性能下降的主要物理机制是支撑膜使靶壳生成大幅度的尖钉深入DD气体区,使烧蚀层与DD气体之间物质界面处的电子热传导漏失功率显著增大,导致DD核反应速率显著降低,中子反应速率峰值时刻（bang-time时刻）提前.相比一维理想内爆的模拟结果,支撑膜引入的扰动显著降低bang-time时刻DD气体压强与内爆动能转化为DD气体内能的效率,壳层剩余动能相应大幅增加.     

Modified Gerchberg–Saxton (G-S) Algorithm and Its Application

The Gerchberg–Saxton (G-S) algorithm is a phase retrieval algorithm that is widely used in beam shaping and optical information processing. However, the G-S algorithm has difficulty obtaining the exact solution after iterating, and an approximate solution is often obtained. In this paper, we propose a series of modified G-S algorithms based on the Fresnel transform domain, including the single-phase retrieval (SPR) algorithm, the double-phase retrieval (DPR) algorithm, and the multiple-phase retrieval (MPR) algorithm. The analysis results show that the convergence of the SPR algorithm is better than that of the G-S algorithm, but the exact solution is not obtained. The DPR and MPR algorithms have good convergence and can obtain exact solutions; that is, the information is recovered losslessly. We discuss the security advantages and verification reliability of the proposed algorithms in image encryption. A multiple-image encryption scheme is proposed, in which n plaintexts can be recovered from n ciphertexts, which greatly improves the efficiency of the system. Finally, the proposed algorithms are compared with the current phase retrieval algorithms, and future applications are discussed. We hope that our research can provide new ideas for the application of the G-S algorithm.     

Visual Speech Recognition with Lightweight Psychologically Motivated Gabor Features

Extraction of relevant lip features is of continuing interest in the visual speech domain. Using end-to-end feature extraction can produce good results, but at the cost of the results being difficult for humans to comprehend and relate to. We present a new, lightweight feature extraction approach, motivated by human-centric glimpse-based psychological research into facial barcodes, and demonstrate that these simple, easy to extract 3D geometric features (produced using Gabor-based image patches), can successfully be used for speech recognition with LSTM-based machine learning. This approach can successfully extract low dimensionality lip parameters with a minimum of processing. One key difference between using these Gabor-based features and using other features such as traditional DCT, or the current fashion for CNN features is that these are human-centric features that can be visualised and analysed by humans. This means that it is easier to explain and visualise the results. They can also be used for reliable speech recognition, as demonstrated using the Grid corpus. Results for overlapping speakers using our lightweight system gave a recognition rate of over 82%, which compares well to less explainable features in the literature.     

电枢参数对电枢超导型高温超导电机性能的影响

高温超导材料存在背景磁场依赖性和各向异性的问题,针对超导感应电机采用双层绕组结构,提出了双重镜像法建立开口矩形槽内漏磁场垂直方向和水平方向的解析模型,利用该模型研究了超导线圈在传输不同电流情况下磁场分布规律,揭示了开口矩形槽电磁结构参数对槽内超导线圈所受表面磁场的影响规律。研究结果表明,适当增大槽宽尺寸可以有效降低超导线圈表面所受最大磁场,传输电流和线圈匝数均能够近乎线性增大超导线圈表面所受最大磁场。     

Real time addition or subtraction of two images by vector wave interferometry

The present article reports a new optical method for real time addition or subtraction of two photographic images. Two transparencies which are placed in the two arms of a Mach-Zehnder interferometer with two orthogonal polarizers are subsequently added or subtracted by rotating an analyser inserted in the output beam. A linear combination of the amplitudes of the two images can also be observed for different orientations of the transmission axis of the analyser. Some experimental results are presented.     

利用三色光栅实现彩色图像相减

本文提出了一种彩色图像相减的新方法.该方法利用互相错动半个周期的两组三色光栅,在黑白底片上分别对待处理的两幅图像进行一次编码记录,然后在白光处理器上对黑白编码片进行彩色译码,即可实现彩色图像的强度相减,文中讨论了三色光栅实时生成和两组三色光栅条纹自动错动半个周期的方法,并给出了实验结果。     

用维纳滤波法复原闪光X光图片

 本文对某系统拍摄的闪光X光模糊图片成功地进行了复原处理，把闪光X光照相系统简化为线性位移不变系统，建立了成像过程的数学模型；系统的退化效应用综合点扩展函数描述，并用针孔和台阶照相的方法进行测量；提出了校正图片上数据非线性及削弱图片中噪声的方法；采用维纳滤波方法作复原运算，给出了复原结果。     

爆炸推进层靶的理论模型

本文通过对爆炸推进层靶的一系列实验结果的分析,提出了一个简化的理论模型,并介绍了估计DT压缩比、离子温度、热核反应约束时间及中子产额的某些方法。在中子产额大于10~6时,计算结果与实验测置值在2—3倍之内符合。