Single mode optical fiber coupling to a laser diode

A double ball lens coupling scheme has been designed. The efficiency of the coupling has been tested. The effect of the lateral and transversal offset on the coupling efficiency has been investigated. Two spherical lenses based on As₂S₃ and As₂Se₃ have been used to this purpose. We used a Roithner RLT7605MG 5 mW single mode laser diode as source. The variation of the coupling with the working distance has been measured and an optimum separation distance has been found. The article is published in the original.     

Design of low power,high speed 4 bit binary to Gray converter with 8 × 4 barrel shifter using nano dimensional MOS transistor for arithmetical,logical and telecommunication circuit and system application

In this work the design of 4 bit binary to Gray code converter circuit with 8 × 4 barrel shifter has been carried out. The circuit has been designed using metal oxide semiconductor (MOS) transistor. The verification of the functionality of the circuits has been performed using Tanner-SPICE software. Power consumption and speed are the major design metrics for very large scale integrated circuit. In this work the average power consumption and gate delay analysis of 4 bit binary to Gray converter with 8 × 4 barrel shifter has been carried out using nano dimensional MOS transistor having channel length of 150 nm. Power consumption, delay analysis has been carried out for different set of supply voltage. It has been observed that power consumption of the 4 bit binary to Gray converter with 8 × 4 barrel shifter has been reduced by reducing the power supply voltage V_DD. The power consumption and delay offers by the circuit is very less. At 1 V V_DD, power consumption and delay are 0.15 μW and 52.7 ps respectively. Therefore the circuit is suited for low power and high speed application in the area of arithmetical, logical and telecommunication.

     

Design, fabrication and characterization of an externally actuated ON/OFF microvalve

A magnetic microfluidic valve has been designed, fabricated and tested. Operation relies on the use of a permanent magnet which interacts with an electrodeposited layer of Co–Ni (soft magnetic material) on a V-shaped cantilever beam. The deflection caused by the magnetic forces opens or closes the fluid flow. The microvalve performance has been optimized by means of finite element analysis (FEA). The FEA model has been experimentally validated using confocal microscopy and used to improve the magnetic circuit. Then, a fluidic cell has been built and the microvalve has been demonstrated to work as a check-valve or as ON/OFF valve when being magnetically actuated. Fabricated prototypes were evaluated in a flow of N₂at the flow rate of 20 sccm. The operational applied pressure was 50 mbar. The microvalve has a leaking rate in the order of 1.75 sccm at 50 mbar.     

LRUNet:轻量级脑肿瘤快速语义分割网络

目的 针对目前基于深度学习的脑肿瘤分割算法参数量大、计算复杂和快速性差的问题,提出了一种超轻量级快速语义分割网络LRUNet （lightweight rapid UNet）,在保证分割精度提升的同时,极大地减少了网络的参数量与计算量,达到快速分割的效果。方法 LRUNet网络结构基于UNet,将3D-UNet的通道数减少为原来的1/4,减少原先3D-UNet过多的参数量;将UNet网络中除最后一层外的所有传统卷积变为深度可分离卷积,深度可分离卷积以牺牲极少精度,大大减少网络参数量,实现网络的轻量级;使用空间—通道压缩和激发模块（spatial and channel squeeze&excitation block,scSE）,该模块能够放大特征图中对模型有利的参数的权重,缩小对模型不利参数的权重,提升网络分割的精度。结果 在BraTS 2018（Brain Tumor Segmentation Challenge 2018）数据集上的在线验证结果显示,该模型在全肿瘤、核心区肿瘤和增强区肿瘤分割的平均Dice系数分别为0.893 6、0.804 6和0.787 2。LRUNet与同为轻量级网络的S3D-UNet相比Dice有所提升,但是,参数量仅为S3D-UNet的1/4,FLOPs （floating point operations per second）仅为1/2。结论 与3D-UNet、S3D-UNet和3D-ESPNet等算法相比,LRUNet算法不仅保证精度得到提升,而且极大地减少网络中计算的参数量与计算成本消耗,同时网络模型的预测速度得到很大提升,使得快速语义分割在3维医学图像领域成为可能。     

轻量级注意力特征选择循环网络的超分重建

目的 深度卷积网络在图像超分辨率重建领域具有优异性能,越来越多的方法趋向于更深、更宽的网络设计。然而,复杂的网络结构对计算资源的要求也越来越高。随着智能边缘设备（如智能手机）的流行,高效能的超分重建算法有着巨大的实际应用场景。因此,本文提出一种极轻量的高效超分网络,通过循环特征选择单元和参数共享机制,不仅大幅降低了参数量和浮点运算次数（floating point operations,FLOPs）,而且具有优异的重建性能。方法 本文网络由浅层特征提取、深层特征提取和上采样重建3部分构成。浅层特征提取模块包含一个卷积层,产生的特征循环经过一个带有高效通道注意力模块的特征选择单元进行非线性映射提取出深层特征。该特征选择单元含有多个卷积层的特征增强模块,通过保留每个卷积层的部分特征并在模块末端融合增强层次信息。通过高效通道注意力模块重新调整各通道的特征。借助循环机制（循环6次）可以有效提升性能且大幅减少参数量。上采样重建通过参数共享的上采样模块同时将浅层与深层特征进放大、融合得到高分辨率图像。结果 与先进的轻量级网络进行对比,本文网络极大减少了参数量和FLOPs,在Set5、Set14、B100、Urban100和Manga109等基准测试数据集上进行定量评估,在图像质量指标峰值信噪比（peak signal to noise ratio,PSNR）和结构相似性（structural similarity,SSIM）上也获得了更好的结果。结论 本文通过循环的特征选择单元有效挖掘出图像的高频信息,并通过参数共享机制极大减少了参数量,实现了轻量的高质量超分重建。     

Block-matching-based motion field generation utilizing directional edge displacement

A motion field generation algorithm using block matching of edge flag histograms has been developed aiming at its application to motion recognition systems. Use of edge flags instead of pixel intensities has made the algorithm robust against illumination changes. In order to detect local motions of interest effectively, a new adaptive frame interval adjustment scheme has been introduced in which only the edge flags due to local motions present in the frame are accumulated and utilized in block matching. These edge flags are projected onto x and y axes to generate histograms and the motion in x and y directions are determined by histogram matching. As a result, the computational cost for best match search has been substantially reduced. A vector representation of the motion field, which is called projected principal-motion distribution (PPMD), has also been proposed. It was applied to preliminary motion recognition experiments using Hidden Markov Models (HMMs) and its effectiveness has been confirmed. Moreover the advantage of the proposed motion field generation method over the simple optical flow as well as the conventional block matching method using pixel intensities has been demonstrated.     

Identification of actors drawn in Ukiyoe pictures

This paper presents the development of line image keywords for the identification of actors drawn in Japanese traditional painting pictures known as Ukiyoe pictures. The system is based on visual features of the face from the image database files and is organized as a set of classifiers whose outputs are integrated after a normalization step. Line profile from the picture has been extracted in this investigation and has been approximated by Bézier curves. A learning algorithm has been developed to obtain the control points at high accuracy. A new curve matching method has been developed based on the feature points, rather than the corresponding points. This method can automatically fit a set of data points with piecewise geometrically continuous third order Bézier curves. Last of all, a new approach for distance calculation, namely “apple-node distance” has been introduced here for similarity calculation in image retrieval systems. The computation of similarity between curves has been established on the basis of this “apple-node” distance. The effectiveness of our method has been confirmed through computer simulation. The method developed here can be expanded to one of three dimensional shape-analyzing tools.     

DARSS: a hybrid mesh smoother for all hexahedral meshes

A method for smoothing hexahedral meshes has been developed. The method consists of two phases. In the first phase, the nodes are moved based on an explicit formulation. A constraint has also been implemented to prevent the deterioration of elements associated with the node being moved. The second phase of the method is optismoothing based on the Nelder–Mead simplex method. The summation of the Jacobian of all the elements sharing a node has been taken as the function to be maximized. The method has been tested on meshes up to 18,305 hexahedral elements and was found to be stable and improved the mesh in about 112.6 s on an Intel Centrino^? 1.6 GHz, 1 GB RAM machine. The method thus has the advantage of being effective as well as being computationally efficient.     

Dynamic characteristics and forced response of an electrostatically-actuated microbeam subjected to fluid loading

In this paper flexural vibrations of an electrostatically actuated cantilever microbeam in an incompressible inviscid stationary fluid have been studied. By applying “Three dimensional aerodynamic theory” pressure jump across the microbeam has been investigated and the inertial effects of fluid on microbeam dynamics have been modeled as a mass added to microbeam mass. Magnitude of the added mass has been calculated for various aspect ratios of cantilever microbeams and compared with those of clamped-clamped microbeams. To investigate the dynamic characteristics, it has been considered that the microbeam has been deflected by a DC voltage, V _DC and then the dynamic characteristics and forced response of the system have been considered about these conditions. Galerkin-based step by step linearization method (SSLM) and Galerkin-based reduced order model have been applied to solve the nonlinear static and dynamic governing equations, respectively. Water by neglecting viscidity effects, as an instant has been considered as a surrounding fluid and the frequency response of the microbeam has been compared with that of vacuum conditions. It has been shown that because of the added mass effects in watery environment, the natural frequencies of the microbeam decrease. Because of the higher dielectric coefficient and increasing electrical stiffness and decreasing total stiffness consequently, maximum amplitude of the microbeam vibrations increases in watery environment, compared with vacuum. Moreover, it has been shown that increasing the DC voltage, increases the electrical stiffness and maximum amplitude of the microbeam vibrations, consequently, It has been shown that in higher voltages (near pull-in voltage), the rate of variation of resonance frequency and maximum amplitude is stronger than lower voltages.     

Rotation and intensity invariant shoeprint matching using Gabor transform with application to forensic science

Shoe marks at the place of crime provide valuable forensic evidence. This paper presents a technique for rotation and intensity invariant automatic shoeprint matching. Multiresolution features of a shoeprint have been extracted using Gabor transform. Rotation of the shoeprint image has been estimated using Radon transform and is compensated by rotating the features in opposite direction. The performance of the proposed algorithm has been compared with the technique in which the features have been determined using Fourier transform and its power spectral density. Shoeprint database has been generated by inviting participants to tread on an inkpad and then stamp on a piece of paper. Euclidian distance classifier has been used to find a suitable match. The performance of the proposed algorithm has been evaluated in terms of correct recognition rate computed using best match score at rank ‘1’ and cumulative match score for the first four matches with rotation, intensity and/or mixed attacks. A good matching performance has been achieved with rotation attack; typically 91 percent at rank ‘1’ and 100 percent at rank ‘2’ for full prints. Performance of the proposed technique is better even for partial shoeprints. Experimentation has also been carried out by perturbing shoeprint images with Gaussian white noise, salt and pepper noise to evaluate the robustness of the proposed technique.     

Monotonic and nonmonotonic gentzen deduction systems for L3-valued propositional logic

A sequent is a pair (Γ, Δ), which is true under an assignment if either some formula in Γ is false, or some formula in Δ is true. In L₃-valued propositional logic, a multisequent is a triple Δ|Θ|Γ, which is true under an assignment if either some formula in Δ has truth-value t, or some formula in Θ has truth-value m, or some formula in Γ has truth-value f. Correspondingly there is a sound and complete Gentzen deduction system G for multisequents which is monotonic. Dually, a comultisequent is a triple Δ : Θ : Γ, which is valid if there is an assignment v in which each formula in Δ has truth-value≠t, each formula in Θ has truth-value≠m, and each formula in Γ has truth-value≠f. Correspondingly there is a sound and complete Gentzen deduction system G⁻ for co-multisequents which is nonmonotonic.     

Curve subdivision schemes for geometric modelling

A new binary four-point approximating subdivision scheme has been presented that generates the limiting curve of C ¹ continuity. A global tension parameter has been introduced to improve the performance of the binary four-point approximating subdivision scheme that generates a family of C ¹ limiting curves. The ternary four-point approximating subdivision scheme has also been introduced that generates a limiting curve of C ² continuity. The proposed schemes are close to being interpolating. The Laurent polynomial method has been used to investigate the order of derivative continuity of the schemes and Hölder exponents of the schemes have also been calculated. Performances of the subdivision schemes have been exposed by considering several examples.     

融合环境特征与改进YOLOv4的安全帽佩戴检测

目的 在施工现场，安全帽是最为常见和实用的个人防护用具，能够有效防止和减轻意外带来的头部伤害。但在施工现场的安全帽佩戴检测任务中，经常出现难以检测到小目标，或因为复杂多变的环境因素导致检测准确率降低等情况。针对这些问题，提出一种融合环境特征与改进YOLOv4（you only look once version 4）的安全帽佩戴检测方法。方法 为补充卷积池化等过程中丢失的特征，在保证YOLOv4得到的3种不同大小的输出特征图与原图经过特征提取得到的特征图感受野一致的情况下，将两者相加，融合高低层特征，捕捉更多细节信息；对融合后的特征图采用3×3卷积操作，以减小特征图融合后的混叠效应，保证特征稳定性；为适应施工现场的各种环境，利用多种数据增强方式进行环境模拟，并采用对抗训练方法增强模型的泛化能力和鲁棒性。结果 提出的改进YOLOv4方法在开源安全帽佩戴检测数据集（safety helmet wearing dataset，SHWD）上进行测试，平均精度均值（mean average precision，mAP）达到91.55%，较当前流行的几种目标检测算法性能有所提升，其中相比于YOLOv4，mAP提高了5.2%。此外，改进YOLOv4方法在融合环境特征进行数据增强后，mAP提高了4.27%，在各种真实环境条件下进行测试时都有较稳定的表现。结论 提出的融合环境特征与改进YOLOv4的安全帽佩戴检测方法，以改进模型和数据增强的方式提升模型准确率、泛化能力和鲁棒性，为安全帽佩戴检测提供了有效保障。     

Understanding human behavior from motion imagery

Computer vision is gradually making the transition from image understanding to video understanding. This is due to the enormous success in analyzing sequences of images that has been achieved in recent years. The main shift in the paradigm has been from recognition followed by reconstruction (shape from X) to motion-based recognition. Since most videos are about people, this work has focused on the analysis of human motion. In this paper, I present my perspective on understanding human behavior. Automatically understanding human behavior from motion imagery involves extraction of relevant visual information from a video sequence, representation of that information in a suitable form, and interpretation of visual information for the purpose of recognition and learning about human behavior. Significant progress has been made in human tracking over the last few years. As compared with tracking, not much progress has been made in understanding human behavior, and the issue of representation has largely been ignored. I present my opinion on possible reasons and hurdles for slower progress in understanding human behavior, briefly present our work in tracking, representation, and recognition, and comment on the next steps in all three areas.Published online: 28 August 2003     

Visual tracking using convolutional features with sparse coding

Artificial Intelligence Review - Visual object tracking has become one of the most active research topics in computer vision, and it has been applied in several commercial...     

EuLisp in education

We present our experience withEuLisp as a teaching language, focussing on the level of the language which was specifically designed for this purpose (level-0).EuLisp has been used in undergraduate and postgraduate teaching since 1990, in lectures and laboratories, where in many cases it has replaced Scheme or Common Lisp. It has been used extensively in programming courses, parallelism courses, as a vehicle for advanced courses in symbolic computing and programming language design; it has also been used as a platform for final year undergraduate projects. This experience has demonstrated thatEuLisp is well suited to teaching and far reaching in its capabilities: it supports the relevant concepts in a consistent and versatile framework, so that the language serves to facilitate the educational process. The discussion is illustrated with examples, and where appropriate we draw a comparison with the Lisp dialects used previously in these courses.     

The bounds of min-max pair heap construction

In this paper, lower and upper bounds for min-max pair heap construction has been presented. It has been shown that the construction of a min-max pair heap with n elements requires at least 2.07n element comparisons. A new algorithm for creating min-max pair heap has been devised that lowers the upper bound to 2.43n.     

A personal conditioning system based on a liquid-conditioned vest and a thermoelectric supply system.

Abstract

A personal conditioning system has been developed for RAF air crew. It employs a liquid-conditioned vest covering the torso and upper arms. The liquid supply system uses thermoelectric elements to provide either heating or cooling. The system has been subjected to a number of environmental tests. It has been shown to be capable of extracting 150 W from a human subject in an ambient temperature of 40^°C and at altitudes of 10000 feet. Further physiological tests have shown that this is sufficient to control body temperature in environments where heat stress would otherwise have occurred. A series of instrumented flights has been carried out in a fighter aircraft. The system has performed successfully and is under consideration for use in future aircraft.     

An economical tonal display for interactive graphics and image analysis data

The use of tonal displays in image analysis and interactive graphics has always dictated the use of expensive refresh memories for the display output device. This has involved the use of high speed digital drums, multiple head discs, and analog storage tubes. Recently, the introduction of very long shift registers has allowed the designer to consider their use for refresh memories. A prototype display using 1024 bit MOS static shift registers has been developed. It has been shown that a reasonable cost versus performance tradeoff can be obtained. The first efforts has resulted in a 128 × 128 × 4 bit (64k) memory; it is now in the process of being expanded to 256 × 256 × 8 bits (512k). This memory is cost competitive with digital disc memories and both cost and performance competitive with storage tube scan converters.     

区域级通道注意力融合高频损失的图像超分辨率重建

目的 通道注意力机制在图像超分辨率中已经得到了广泛应用,但是当前多数算法只能在通道层面选择感兴趣的特征图而忽略了空间层面的信息,使得特征图中局部空间层面上的信息不能合理利用。针对此问题,提出了区域级通道注意力下的图像超分辨率算法。方法 设计了非局部残差密集网络作为网络的主体结构,包括非局部模块和残差密集注意力模块。非局部模块提取非局部相似信息并传到后续网络中,残差密集注意力模块在残差密集块结构的基础上添加了区域级通道注意力机制,可以给不同空间区域上的通道分配不同的注意力,使空间上的信息也能得到充分利用。同时针对当前普遍使用的L1和L2损失函数容易造成生成结果平滑的问题,提出了高频关注损失,该损失函数提高了图像高频细节位置上损失的权重,从而在后期微调过程中使网络更好地关注到图像的高频细节部分。结果 在4个标准测试集Set5、Set14、BSD100（Berkeley segmentation dataset）和Urban100上进行4倍放大实验,相比较于插值方法和SRCNN（image super-resolution using deep convolutional networks）算法,本文方法的PSNR（peak signal to noise ratio）均值分别提升约3.15 dB和1.58 dB。结论 区域级通道注意力下的图像超分辨率算法通过使用区域级通道注意力机制自适应调整网络对不同空间区域上通道的关注程度,同时结合高频关注损失加强对图像高频细节部分的关注程度,使生成的高分辨率图像具有更好的视觉效果。