Abstract: An implementation cost of a line-scan optical coherence tomography (OCT) apparatus is reduced by miniaturizing a scanning mirror and using a light source with relaxed requirement in intensity uniformity. The mirror reflects a probe light beam to different parts of a sample for line-scanning the sample. A line-compressing lens compresses the probe light beam's cross-sectional length before the beam reaches the mirror, allowing the mirror to be miniaturized to reflect only the compressed beam. In generating a linear light beam that gives the probe light beam, a cascade of collimating lens, Powell lens and focusing lens generates the linear light beam from a raw light beam of a point source. A slit further filters the linear light beam to remove a peripheral portion thereof such that the linear light beam is substantially uniform in intensity even if an asymmetrical divergent light source is used.

Abstract: A hardware accelerator for accelerating the zero knowledge succinct non-interactive argument of knowledge (zk-SNARK) protocol by reducing the computation time of the cryptographic verification is disclosed. The accelerator includes a zk-SNARK engine having one or more processing units running in parallel. The processing unit can include one or more multiply-accumulate operation (MAC) units, one or more fast Fourier transform (FFT) units; and one or more elliptic curve processor (ECP) units. The one or more ECP units are configured to reduce a bit-length of a scalar di in an ECP algorithm used for generating a proof, thereby the cryptographic verification requires less computation power.

Abstract: A method for a protocol in digital asset trading includes step as follow. Trading information is exchanged between the first and the second nodes, in which a first digital currency belonging to the first node and a second digital currency belonging to the second node are expected to be exchanged, and prices of the first and second digital currencies are time-dependent. An atomic swap script and a smart contract are generated. A compensation fee to be paid to the second node by the first node is computed. A redeeming or refunding event is performed in response to the atomic swap script and the smart contract with paying the compensation fee to the second node, in which the compensation fee is dependent on the prices of the first and second digital currencies such that is time-dependent.

Abstract: An apparatus for managing data messages comprises: one or more producers generating data streams containing data messages of varying sizes that required processing; one or more consumers for processing the data messages; a multi-message queues sub-system for queuing data messages having different processing time durations; a rate limiter for discriminating data messages based on processing speed for queuing the data messages in one or the other message queue of the multi-message queues sub-system; a fair dispatcher for dispatching the data messages to one or more consumers according to their processing statuses to maximize the processing capacity of the apparatus; and a task splitter for splitting data messages that are deemed too large.

Abstract: A dynamic-tile neural network accelerator allows for the number and size of computational tiles to be re-configured. Each sub-array of computational cells has edge cells on the left-most column that have an added vector mux that feeds the cell output back to an adder-comparator to allow Rectified Linear Unit (ReLU) and pooling operations that combine outputs shifted in from other cells. The edge cells drive external output registers and receive external weights. The weights and outputs are shifted in opposite directions horizontally between cells while control and input data are shifted in a same direction vertically between cells. A column of row data selectors is inserted between sub-arrays to bypass weights and output data around sub-arrays, while a row of column data selectors are inserted between sub-arrays to bypass control and input data. Larger tiles are configured by passing data directly through these selectors without bypassing.

Abstract: A method of high dimensional data analysis in real-time comprising executing dimension-reducing an input historical data set under a t-SNE model and determining from the resulting dimension-reduced data set a recent; further dimension-reducing the recent group data set under a PCA model; statistical analyzing the further dimension-reduced data set to determine a threshold group for distinguishing abnormal data from normal ones in a real-time data stream. The method may further include training a classifier using the abnormal or normal data set for predicting anomaly in the real-time data source system. Alternatively, a discrepancy training data set is computed from one of the normal and abnormal data sets and be used to train one of independent normal and abnormal data regression models; with the other one trained by transfer learning based on the trained one. The trained regression models are then used to predict discrepancy values.

Abstract: A module has electronic components mounted to a Printed Circuit Board (PCB) with multiple patterned conductive layers connecting to conductive slot metal around a conductive slot. A groove is cut through a top molding encapsulant above and into the conductive slot but does not cut through a bottom molding encapsulant. A terminal pin is inserted into the groove and pushed down into the conductive slot. When heated, embedded solder previously applied to the conductive slot metal flows between the end of the terminal pin and the conductive slot metal to form a solder bond. An end of the PCB past the conductive slot has no metal traces, preventing shorts. Epoxy can be placed into the groove around the terminal pin or a hole formed in the terminal pin to increase strength of the anchored terminal pin. The molding around the groove protects terminal pins from shorting from the side.

Abstract: A neural network receives mono-color pixels in a Bayer pattern from an image sensor and interpolates pixels to generate full-color RGB pixels while also enlightening the image for better detail. A Back-Projection (BP) layer is added to each contracting layer in a U-net convolution neural network where feature depth is increased as pixels are downsampled, while a BP channel shrink layer is added to each expansion layer where feature depth is reduced to upsample and increase pixel resolution. Convolution layers in the BP layer lighten and darken images to generate an error that is then lightened and added to a weighted lightened input. The neural network learns the best weightings to correct for noise and error in the lightening process in these BP layers. Noise is reduced further by repeating a convolution and leaky Rectified Linear Unit (lrelu) in series for the first convolution in the BP layer that performs lightening.

Abstract: A used battery is discharged for a short time from a first Open Circuit Voltage (OCV1) to a second OCV2 and the discharge current ?Q measured. OCV1 is input to a calibration curve model to obtain a first modeled State of Charge (SOC1) value, and OCV2 is input to the calibration curve model to obtain a second modeled SOC2 value. The State of Health (SOH) is calculated as ?Q/[Qnew×(SOC1?SOC2)], where Qnew is the battery capacity when new. The used battery is sorted for reuse or disposal based on the SOH value. The calibration curve model is obtained by Artificial Intelligence (AI) modeling of OCV, SOC datapoints from fully charging and discharging used batteries. Only OCV values within a target region having a low first derivative of SOC as a function of OCV are modeled, and OCV1 and OCV2 are within this target region.

Abstract: A forensic detector for remotely sensing a porphyrin-containing substance on a surface of an object is provided. The forensic detector includes an excitation light source for generating an excitation light beam, plural optical elements, one or more sensors for measuring a first and second radiation powers, and a computer processor. The excitation light beam has an excitation wavelength capable of exciting porphyrin to luminesce at a target wavelength centered between 665-675 nm. The first and second radiation powers are in non-overlapping wavelength regions of the object-originated light beam. The computer processor is configured to compute a SNR according to the first radiation power and a background noise power level, and to determine if the porphyrin-containing substance is present on the surface according to the SNR. The background noise power level is obtained from the one or more second radiation powers.

Abstract: A smart power hub has an AC port, a first DC port to a car battery, a second DC port to DC devices such as car instruments, and a third DC port to solar panels or another smart power hub. An AC bi-directional converter has six transistors in a B6 configuration to convert three-phase AC, acting as an interleaved totem-pole Power-Factor-Corrector for one-phase AC. A switch connects the DC bi-directional converter with either the AC bi-directional converter or the solar panels on the third DC port. A link capacitor has a DC link voltage that rises during battery charging. A DC bi-directional converter has a transformer, a primary bridge connected to the DC link voltage, and a secondary bridge of transistors connected to the first DC port. Auxiliary windings in the transformer drive a rectifier to the second DC port to power on-board DC devices. Solar DC charges the battery without AC conversion.

Abstract: A method for extracting information from a table includes steps as follows. Characters of a table are extracted. The characters are merged into n-gram characters. The n-gram characters are merged into words and text lines through a two-stage GNN mode. The two-stage GNN mode comprises sub steps as: spatial features, semantic features, CNN image features are extracted from a target source; a first GNN stage is processed to output graph embedding spatial features from the spatial features; and a second GNN stage is processed to output graph embedding semantic features and graph embedding CNN image features from the semantic features and the CNN image features, respectively. The text lines are merged into cells. The cells are grouped into rows, columns, and key-value pairs based on one or more adjacency matrices, a row relationship among the cells, a column relationship among the cells, and a key-value relationship among the cells.

Abstract: Computer implemented methods and computerized apparatus for identifying planes of a 3D object in a 3D scene are provided. The method comprises receiving a point cloud and performs plane segmentation. In one embodiment, the point cloud is voxelized into a plurality of voxels of equal dimensions and the plurality of voxels are classified into three categories, wherein each first category voxel satisfies a planar requirement and a first neighborhood constraint, each second category voxel satisfies the planar requirement and a second neighborhood constraint, and each third category voxel does not satisfy the planar requirement.

Abstract: A wireless power transmitter has electromagnets and a transmitter coil mounted on a moving plate. Current in the main DC circuit is switched into the electromagnets by transistors that control the amount and direction of current through the electromagnets to control the strength and polarity of electromagnetic fields generated by the electromagnets. A controller initially drives a high gate voltage onto transistors to cause the electromagnets on the transmitter to generate a maximum attractive force with magnets on the receiver, causing the moving plate to move the transmitter coil into closer alignment with the receiver coil. A power factor is measured on both receiver and transmitter to estimate the power transfer efficiency. The controller then reduces the gate voltage by a step size and the power factors are measured again. The gate voltage continues to be adjusted to optimize the power transfer efficiency until reaching a maxima.

Abstract: A method for inter-blockchain communication, comprising: receiving from a sender node in a sender blockchain a data message for transmission to a receiver node in a receiver blockchain; encrypt and encapsulate by an encryption module the data message into an event, wherein the event comprising an exposed header containing information for routing the event through the blockchains and identifying the sender and the receiver nodes; broadcasting the event to a communication bridge comprising multiple bridge nodes; transferring the event through the bridge nodes to an event exchange module of the receiver blockchain; validating the event by a validation module; decrypting and decapsulating the validated event by a decryption module into a decrypted data message; and recording the decrypted data message into the receiver blockchain ledger for reading by the receiver node. The multiple bridge nodes ensure redundancy for the reliable delivery of events in the inter-blockchain communication.

Abstract: A multilayered optical security device has a resonant waveguide grating (RWG) layer on an under layer having a hologram. The RWG layer has plural RWGs for generating one or more patterns. Each RWG has first and second RWG portions, with diffraction gratings of different grating periods, for light coupling-in and coupling-out, respectively. Using the RWGs minimizes interference to a holographic image generated by the under layer. This advantage is especially useful when the holographic image carries digital data such as in digital holographic encryption where an encrypted hologram is used in the under layer. The digital data may contain biometric information such as a fingerprint. In each RWG portion of an individual RWG, scattering elements in a grating are curved to thereby enhance an angular tolerance in azimuthal angle during observing the patterns by an observer while the individual RWG is still selective in elevation angle.

Abstract: A method for estimating camera orientation relative to a ground surface. Line segments are detected from an image captured by a camera. A first virtual cube having three orthogonal vanishing points with a random 3D orientation is superimposed on to the image. The line segments of the image are classified grouped into 3D-directional groups. A second virtual cube is superimposed on to the image with an initial 3D orientation. An optimal 3D orientation of the second virtual cube is computed by iteratively changing the 3D orientation of the second virtual cube and measuring perpendicular distances of the three orthogonal vanishing points to the three line segment groups in each iteration starting with the initial 3D orientation, wherein the optimal 3D orientation of the second virtual cube being one that provides shortest perpendicular distances. Co-variances of the orthogonal vanishing points of the second virtual cube at the optimal orientation are computed.

Abstract: A method for summarizing text with sentence extraction including steps as follows. Sentences are extracted from a document including text by a natural language processing (NLP) based feature extractor. A word vector set with respect to each of the sentences is generated by a processor. The word vector set with respect to each of the sentences is used to generate a n-grams vector set and a phrase-n vector set with respect to each of the sentences. A word score representing similarity between the word vector sets, a n-grams score representing similarity between the n-grams vector sets, and a phrase-n score representing similarity between the phrase-n vector sets are computed. The word, n-grams, and phrase-n scores are combined to compute an edge score. Text features are selected from the sentences using the edge scores of the sentences, so as to output a summary of the document.

Abstract: An adaptive dynamic planning control method and system for a large-scale energy storage station. The method comprises: setting a structure and control target parameters of an adaptive dynamic planning control system; initializing the parameters and importing an initial state of a controlled object; calculating an original wind electricity power fluctuation rate at a current moment t and smoothing the original wind electricity power according to a change rate control strategy; calculating a smoothed wind storage power fluctuation rate, a power of an energy storage system, and a state of charge (SOC) of the energy storage system; initializing and training an evaluation module and an execution module; calculating and saving a control strategy, a smoothed wind storage power fluctuation rate, an energy storage power and a (SOC) at each moment; and outputting the control strategy at each moment, the smoothed wind storage power fluctuation rate, the energy storage power and the (SOC).

Abstract: The present invention provides an anomaly detection method and apparatus based on a neural network which can be trained on undamaged normal vehicle images and able to detect unknown/unseen vehicle damages of stochastic types and extents from images which are taken in various contexts. The provided method and apparatus are implemented with functional units which are trained to perform the anomaly detection under a GCAN model with a training dataset containing images of undamaged vehicles, intact-vehicle frame images and augmented vehicle frame images of the vehicles.