摘要：【中文】本发明提供了一种用于测量电沉积溶液中金属离子浓度的方法。本发明的方法可以实质上减少有机添加剂和不同电极条件对伏安法测量金属离子浓度的干扰。 【EN】The invention provides a method for measuring the concentration of metal ions in an electrodeposition solution. The method of the present invention can substantially reduce interference of organic additives and different electrode conditions on voltammetry measurements of metal ion concentrations.

摘要：【中文】接收器由来自线圈天线的无线功率供电。匹配网络将接收器调谐到谐振频率，桥和电容器产生输出电压。输出电压被分压并与参考电压进行比较。当比较结果为真时，异步数字控制器增加一个数字计数，但当比较结果为假时，减少所述数字计数。电流导向数模转换器(DAC)从输出端分流电流，其是所述数字计数的一个函数。异步数字控制器、比较器和DAC不使用系统时钟，因此到分流电流的数字反馈在达到目标输出电压时工作，在等待系统时钟开始脉冲时防止过压。将所述数字计数与一个数字阈值进行比较，以恢复发射的幅度移位键控(ASK)数据。 【EN】The receiver is powered by wireless power from the coil antenna. The matching network tunes the receiver to the resonant frequency, and the bridge and capacitor produce an output voltage. The output voltage is divided and compared with a reference voltage. The asynchronous digital controller increments a digital count when the comparison is true, but decrements the digital count when the comparison is false. A current-steering digital-to-analog converter (DAC) shunts current from the output, which is a function of the digital count. The asynchronous digital controller, comparator and DAC do not use the system clock, so the digital feedback to the shunt current works when the target output voltage is reached, preventing over-voltage while waiting for the system clock to start pulsing. The digital count is compared to a digital threshold to recover transmitted Amplitude Shift Keying (ASK) data.

摘要：【中文】通过2D卷积神经网络(CNN)分割由多个解剖学图像形成的体积图像，每个解剖学图像具有不同成像模态的多个图像切片。通过在没有任何估计的图像切片的情况下结合来自两个相邻解剖学图像的所选择的图像切片来预处理单个解剖学图像以形成混合上下文图像。2D CNN利用关于多模态上下文的边信息和3D空间上下文来提高分割准确性，同时避免由于估计的图像切片中的伪像导致的分割性能降低。2D CNN通过具有从最高层级到最低层级的多个层级的BASKET‑NET模型来实现。层级的大多数多通道特征图中的通道数量从最高层级到最低层级单调减少，从而允许最高层级富含低层级特征细节，以帮助更精细地分割单个解剖学图像。 【EN】A volumetric image formed from a plurality of anatomical images, each having a plurality of image slices of a different imaging modality, is segmented by a 2D Convolutional Neural Network (CNN). A single anatomical image is pre-processed by combining selected image slices from two adjacent anatomical images without any estimated image slices to form a mixed context image. The 2D CNN utilizes side information about multimodal contexts and 3D spatial context to improve segmentation accuracy while avoiding segmentation performance degradation due to artifacts in the estimated image slices. The 2D CNN is implemented by a BASKET-NET model with multiple tiers from the highest tier to the lowest tier. The number of channels in most multi-channel feature maps of a hierarchy decreases monotonically from the highest hierarchy to the lowest hierarchy, allowing the highest hierarchy to be rich in low-level feature detail to help more finely segment a single anatomical image.

摘要：【中文】一种通过句子提取来概括文本的方法，包括以下步骤：通过基于自然语言处理(NLP)的特征提取器从包含文本的文档中提取句子。由处理器关于每个句子生成一个单词向量集。使用关于每个句子的单词向量集生成关于每个句子的n元向量集和短语‑n向量集。计算代表单词向量集之间相似度的单词分数、代表n元向量集之间相似度的n元分数、和代表短语‑n向量集之间相似度的短语‑n分数。单词、n元和短语‑n分数被组合以计算边缘分数。使用句子的边缘分数从句子中选择文本特征，以输出文档的概要。 【EN】A method of summarizing text by sentence extraction, comprising the steps of: sentences are extracted from documents containing text by a Natural Language Processing (NLP) based feature extractor. A set of word vectors is generated by the processor for each sentence. A set of word vectors for each sentence is used to generate a set of n-grams and a set of phrase-n-vectors for each sentence. A word score representing the degree of similarity between the set of word vectors, an n-gram score representing the degree of similarity between the set of n-gram vectors, and a phrase-n score representing the degree of similarity between the phrase-n-gram sets are calculated. Word, n-gram and phrase-n scores are combined to compute an edge score. Text features are selected from the sentence using the edge scores of the sentence to output a summary of the document.

摘要：【中文】一种用于在通信网络中通过高延迟网络连接传输数据包的隧道连接的建立方法，所述方法包括以下步骤：在高延迟网络连接的隧道连接发起方侧配置的模块或节点处，使用非TCP连接来封装建立连接消息，以发送到在回程网络的核心网络(CN)侧上配置的模块或节点，以在所述高延迟网络连接的发起方侧和所述高延迟网络连接的核心网络(CN)侧之间建立数据包传输隧道连接。 【EN】A method of establishing a tunnel connection for transmitting data packets over a high latency network connection in a communication network, the method comprising the steps of: at a module or node configured on a tunnel connection initiator side of a high latency network connection, an establish connection message is encapsulated using a non-TCP connection for transmission to a module or node configured on a Core Network (CN) side of a backhaul network to establish a packet transport tunnel connection between the initiator side of the high latency network connection and the Core Network (CN) side of the high latency network connection.

摘要：【中文】屏蔽型肖特基异质结功率晶体管由碳化硅(SiC)晶圆制成，该晶圆具有SiC外延层，该外延层包括N+源极和位于栅极下方的硅N外延层，其沟道迁移率比SiC高。晶片的块体是与背面金属接触的N+SiC漏极。沟槽形成在异质结晶体管之间。与N+源极接触的金属延伸到沟槽中，与N‑SiC衬底形成肖特基二极管。沟槽侧面的P+抽头将金属连接到异质结栅极下方的P‑SiC体二极管，并防止肖特基金属直接接触P体二极管。在沟槽肖特基二极管下方和异质结晶体管下方形成带有P+柱帽的掩埋P柱。P柱通过与N衬底平衡电荷来提供屏蔽，充当电介质以减小柱上方的电场。 【EN】The shielded schottky heterojunction power transistor is made from a silicon carbide (SiC) wafer having a SiC epitaxial layer comprising an N + source and a silicon N epitaxial layer located below the gate with a higher channel mobility than SiC. The bulk of the wafer is the N + SiC drain in contact with the backside metal. Trenches are formed between the heterojunction transistors. Metal in contact with the N + source extends into the trench forming a schottky diode with the N-SiC substrate. The P + taps at the sides of the trench connect the metal to the P-SiC body diode under the heterojunction gate and prevent the schottky metal from directly contacting the P-body diode. A buried P-pillar with a P + pillar cap is formed under the trench Schottky diode and under the heterojunction transistor. The P pillars provide shielding by balancing charge with the N substrate, acting as a dielectric to reduce the electric field above the pillars.

摘要：【中文】单温点温度敏感传感器假设所有灵敏度线在绝对零温度下会聚，因此在校准期间仅需要在一个温度下进行测量。传感器输出电压(VTS)由来自镜像电流源(28)的电流产生，其流过一个可变电阻器(38)。在校准期间，调节可变电阻器(38)的电阻(R2)和镜像电流源(28)的镜像比。误差放大器(20)比较由单位电流源(22，26)产生的单位电流产生的电压，以调整单位电流源(22，26)和镜像电流源(28)。每个单位电流流过接地的PNP晶体管(31，32)。可开关的PNP晶体管(33)与一个基极接地PNP晶体管(32)并联，其基极接通和断开以调节PNP电流以进行两次测量。将两次测量之间的差值(DVTS')与校准目标(DVTS)进行比较，以在校准期间单一温度下调整可变电阻器(38)和镜像比率。 【EN】A single-point temperature sensitive sensor assumes that all sensitivity lines converge at absolute zero temperature, so only one temperature needs to be measured during calibration. The sensor output Voltage (VTS) is generated by current from a mirror current source (28), which flows through a variable resistor (38). During calibration, the resistance (R2) of the variable resistor (38) and the mirror ratio of the mirror current source (28) are adjusted. An error amplifier (20) compares voltages generated by the unit currents generated by the unit current sources (22, 26) to adjust the unit current sources (22, 26) and the mirror current source (28). Each unit current flows through a PNP transistor (31, 32) connected to ground. The switchable PNP transistor (33) is connected in parallel with a base grounded PNP transistor (32), the base of which is switched on and off to regulate the PNP current for two measurements. The difference between the two measurements (DVTS') is compared to a calibration target (DVTS) to adjust the variable resistor (38) and the mirror ratio at a single temperature during calibration.

摘要：【中文】一种用于跟踪二维平面上移动的目标物体的位置和方向的方法和系统，相机用以生成目标物体的视频帧，处理器被配置以：校正当前一个视频帧的透视比例；使用运动模型或其先前的位置/角度来估计目标物体的当前位置和角度；在估计的当前位置周围从当前视频帧中切出感兴趣的区域；调整感兴趣区域的大小以按预定比例M(宽度＝宽度/M，高度＝高度/M)缩小其尺寸，或等效地以比例M降低感兴趣区域的图像分辨率(分辨率＝分辨率/M)；通过将当前感兴趣区域与先前存储的模板进行模板匹配来跟踪目标物体的新位置；且如果达到了指示成功跟踪的停止标准，则得出结论：所述跟踪的目标物体的位置和角度是准确的，并更新模板；或者继续采用细化方法细化所述跟踪位置和角度，直到最终达到停止标准。 【EN】A method and system for tracking the position and orientation of a moving target object on a two-dimensional plane, a camera to generate video frames of the target object, a processor configured to: correcting the perspective proportion of the current video frame; estimating a current position and angle of the target object using the motion model or its previous position/angle; cutting out a region of interest from the current video frame around the estimated current position; adjusting the size of the region of interest to reduce its size by a predetermined ratio M (width/M, height/M), or equivalently to reduce the image resolution of the region of interest by the ratio M (resolution/M); tracking a new position of the target object by template matching the current region of interest with a previously stored template; and if a stopping criterion indicating successful tracking is reached, it is concluded: the position and angle of the tracked target object are accurate, and the template is updated; or continuing to adopt a thinning method to thin the tracking position and the tracking angle until finally reaching the stopping standard.