Abstract: The disclosure relates to mixed analog-digital circuits, and more specifically a low-noise millimeter-wave fractional-N frequency synthesizer. It overcomes quantization noise and fractional spurs caused by the limited dynamic range and nonlinearity of time error amplifiers (TA) in traditional phase-locked loop structures based on TA. In addition to the traditional structure, the synthesizer includes a coarse digital-to-time converter (CDTC), a fine digital-to-time converter (FDTC), and DTC non-linearity calibration circuits. By inserting the CDTC and FDTC before and after the TA, respectively, the variance of the input phase difference of the TA can be reduced, thereby improving the TA linearity and suppressing the quantization noise and spur generated by fractional-N operation. Furthermore, by using non-linearity calibration, the non-linearity of DTC and TA can be compensated to avoid large quantization noise and spur while the second order quantization noise reshaping is maintained.

Abstract: A single-chain variable fragment (ScFv) derived from anti-c-Met monoclonal antibody MetMAb that specifically binds to c-Met receptor is provided. Also provided are chimeric antigen receptor (CAR) vectors including the ScFv, human T cells transduced with the disclosed CAR vectors, pharmaceutical compositions including the CAR T cells, ScFv fusion proteins, and methods of treating a c-Met-positive cancer or a cancer characterized by overexpression of c-Met in a subject in need thereof by administering an effective amount of the disclosed CAR T cells.

Abstract: This application discloses a scanning electron microscope system to capture an image of an electronic device manufactured according to a layout design describing the electronic device, and a computing system to generate a predicted image of the electronic device using the layout design. The predicted image corresponds to an expected image of the electronic design system captured by the scanning electron microscope system. The computing system identifies manufacturing defects present in the electronic device based on differences between the predicted image of the electronic device and the captured image of the electronic device, and utilizes the captured image of the electronic device to classify the manufacturing defects identified based on the predicted image of the electronic device from the layout design. The computing system can generate a manufacturing defect report identifying the manufacturing defects used to perform repair of the electronic device or modification of the layout design.

Abstract: The disclosure relates to mixed analog-digital circuits, and more specifically a low-noise millimeter-wave fractional-N frequency synthesizer. It overcomes quantization noise and fractional spurs caused by the limited dynamic range and nonlinearity of time error amplifiers (TA) in traditional phase-locked loop structures based on TA. In addition to the traditional structure, the synthesizer includes a coarse digital-to-time converter (CDTC), a fine digital-to-time converter (FDTC), and DTC non-linearity calibration circuits. By inserting the CDTC and FDTC before and after the TA, respectively, the variance of the input phase difference of the TA can be reduced, thereby improving the TA linearity and suppressing the quantization noise and spur generated by fractional-N operation. Furthermore, by using non-linearity calibration, the non-linearity of DTC and TA can be compensated to avoid large quantization noise and spur while the second order quantization noise reshaping is maintained.

Abstract: A data processing method and apparatus, and a device are provided. In this application, a plurality of data processing modules may collaboratively process data. Data output by each data processing module is stored in a data set, the data set includes a plurality of pieces of data, each piece of data carries one index, and the index indicates a data processing module that generates the data. A first data processing module in the plurality of data processing modules may obtain, from the data set, first data carrying a first index, where the first index indicates a data processing module that generates the first data. Then, the first data processing module processes the first data to generate second data carrying a second index, where the second index indicates the first data processing module. Then, the first data processing module stores the second data into the data set.

Abstract: A method for extracting rivet points in large scale three-dimensional point cloud based on deep learning is provided. Geometric attribute scalar of a point cloud of aircraft skin is calculated point by point, and the scalar attribute domain is mapped to the two-dimensional image to obtain a two-dimensional attribute scalar map of the point cloud. The 2D attribute scalar map is processed using a convolutional neural network and the probability that each point belongs to a rivet point is calculated. The rivet point cloud is divided through a threshold according to the probability; and the point clouds belonging to a same rivet is clustered from the divided rivet point cloud using Euclidean cluster.

Abstract: A method for detecting and recognizing surface defects of an automated fiber placement composite based on an image converted from point clouds, including: acquiring a surface point cloud of the automated fiber placement composite; fitting a plane to surface point data; calculating a distance from each point of the surface point cloud to a fitted plane; enveloping the surface point cloud by OBB, and generating a grayscale image according to the OBB and the distance; constructing a pre-trained semantic segmentation network for defect of fiber placement, and inputting the grayscale image to segment and recognize defect areas thereon; mapping a segmentation result output by the semantic segmentation network to the point cloud followed by defect evaluation and visualization.

Abstract: A method for dynamically measuring deformation of a rotating-body mold, including: (S1) subjecting an overall outer surface of the rotating-body mold to three-dimensional measurement to acquire an initial point cloud data; (S2) shooting, by a multi-camera system, the mold from different angles to obtain three-dimensional coordinates of marking points and coding points on the overall outer surface of the rotating-body mold; (S3) rotating the mold, and repeatedly photographing the marking points and the coding points on the mold surface under different angle poses; and calculating three-dimensional coordinates of the marking points and the coding points; and (S4) predicting a point cloud data of the outer surface under different angle poses based on a conversion relationship among the marking points to analyze a deformation degree of the mold during a rotation process.

Abstract: A 3D physical unclonable functions (PUF) system produced based on harnessing the out-of-plane crumpling of a layer of 2D material during shrinkage of a substrate carrying such layer. The structural details of the so-formed 3D PUF pattern are extracted from the tags in a layer-by-layer fashion using confocal laser microscopy imaging and then reconstructed to form the 3D PUF keys and stored in the database, serving as a secure anti-counterfeiting PUF that demonstrates encoding capacity in excess of 1040,000,000. Authentication is performed with a customized trained Siamese neural network framework in a matter of few minutes in a fashion that does not depend on rotation, linear translation, tilt, variations of contrast and/or resolution of the extracted optical images.

Abstract: A method for automated flushness measurement of point cloud rivets, including: extracting a rivet outline by adopting an RANSAC circle fitting algorithm, and determining a center, a radius and a normal vector of an outline circle; extracting point cloud of a rivet head for a single rivet outline; extracting point cloud around the rivet for the single rivet outline; and generating a distance color difference map reflecting rivet flushness according to the point cloud of the rivet head and the point cloud around the rivet. According to the present invention, the point cloud of the rivet head and the point cloud around the rivet can be respectively extracted, and the distance color difference map reflecting the rivet flushness is generated according to the point cloud of the rivet head and the point cloud around the rivet, so that the rivet flushness is rapidly and effectively measured.

Abstract: A method for visualizing a large-scale point cloud based on normal, including: (S1) according to a spatial structure of a point cloud data, constructing a balanced octree structure of a node point cloud; (S2) according to the balanced octree structure and normal information of a point cloud, constructing an octree structure with the normal information; and constructing a normal level-of-detail (LOD) visualization node through downsampling; and (S3) determining a node scheduling strategy according to a relationship between a viewpoint, a viewing frustum and a normal of a render node; and respectively calling a reading thread and a rendering thread to simultaneously perform reading and rendering according to the node scheduling strategy.

Abstract: A method for measuring a seam on aircraft skin based on a large-scale point cloud is disclosed. A point cloud density of each point in an aircraft skin point cloud is calculated. Seam and non-seam point clouds are divided according to a discrepancy of the calculated point cloud density. A point is selected from the point cloud of the seam area, and a section at the point is extracted. A certain range of the seam and non-seam point clouds is projected to the section and a projected point cloud is acquired. A calculation model of flush and gap is constructed, and the flush and the gap of the aircraft skin seam at the measuring point is calculated according to the projected point cloud and the calculation model.

Abstract: A method for automatic glue-spraying of stringers and inspection of glue-spraying quality based on measured data. Three-dimensional (3D) point cloud data of a stringer-skin assembly is collected by 3D laser scanner, and then processed by denoising and sampling. Feature points of an intersection line of a site to be glued of the stringer-skin assembly are extracted by K-means clustering method based on Gaussian mapping, and a minimum spanning tree is constructed based on a set of the extracted feature points. A connected region is established to obtain an initial feature intersection line of the string-skin assembly, which is optimized by random sample consensus algorithm to remove redundant small branch structures to obtain the actual glue-spraying trajectory. The quality of the glue sprayed on the stringer-skin assembly is inspected by line laser to determine positions of the defects, which are then subjected to secondary glue-spraying.

Abstract: A method for visualizing a large-scale point cloud based on normal, including: (S1) according to a spatial structure of a point cloud data, constructing a balanced octree structure of a node point cloud; (S2) according to the balanced octree structure and normal information of a point cloud, constructing an octree structure with the normal information; and constructing a normal level-of-detail (LOD) visualization node through downsampling; and (S3) determining a node scheduling strategy according to a relationship between a viewpoint, a viewing frustum and a normal of a render node; and respectively calling a reading thread and a rendering thread to simultaneously perform reading and rendering according to the node scheduling strategy.

Abstract: A method for detecting and recognizing surface defects of an automated fiber placement composite based on an image converted from point clouds, including: acquiring a surface point cloud of the automated fiber placement composite; fitting a plane to surface point data; calculating a distance from each point of the surface point cloud to a fitted plane; enveloping the surface point cloud by OBB, and generating a grayscale image according to the OBB and the distance; constructing a pre-trained semantic segmentation network for defect of fiber placement, and inputting the grayscale image to segment and recognize defect areas thereon; mapping a segmentation result output by the semantic segmentation network to the point cloud followed by defect evaluation and visualization.

Abstract: A method for dynamically measuring deformation of a rotating-body mold, including: (S1) subjecting an overall outer surface of the rotating-body mold to three-dimensional measurement to acquire an initial point cloud data; (S2) shooting, by a multi-camera system, the mold from different angles to obtain three-dimensional coordinates of marking points and coding points on the overall outer surface of the rotating-body mold; (S3) rotating the mold, and repeatedly photographing the marking points and the coding points on the mold surface under different angle poses; and calculating three-dimensional coordinates of the marking points and the coding points; and (S4) predicting a point cloud data of the outer surface under different angle poses based on a conversion relationship among the marking points to analyze a deformation degree of the mold during a rotation process.

Abstract: A method for extracting rivet points in large scale three-dimensional point cloud based on deep learning is provided. Geometric attribute scalar of a point cloud of aircraft skin is calculated point by point, and the scalar attribute domain is mapped to the two-dimensional image to obtain a two-dimensional attribute scalar map of the point cloud. The 2D attribute scalar map is processed using a convolutional neural network and the probability that each point belongs to a rivet point is calculated. The rivet point cloud is divided through a threshold according to the probability; and the point clouds belonging to a same rivet is clustered from the divided rivet point cloud using Euclidean cluster.

Abstract: A method for measuring a seam on aircraft skin based on a large-scale point cloud is disclosed. A point cloud density of each point in an aircraft skin point cloud is calculated. Seam and non-seam point clouds are divided according to a discrepancy of the calculated point cloud density. A point is selected from the point cloud of the seam area, and a section at the point is extracted. A certain range of the seam and non-seam point clouds is projected to the section and a projected point cloud is acquired. A calculation model of flush and gap is constructed, and the flush and the gap of the aircraft skin seam at the measuring point is calculated according to the projected point cloud and the calculation model.

Abstract: Disclosed herein is a code point-driven three-dimensional (3D) point cloud deformation method. In the method, movable code points and fixed code points are respectively pasted on a moving structure and a static structure. Reference poses of the movable code points and fixed code points are obtained by a dual-camera measurement system, and a 3D point cloud reference model containing the moving structure and the static structure is obtained by 3D laser scanning. A transformation matrix of each code point is calculated, and a real-time point cloud model is established based on the transformation matrix to complete the real-time and dynamic measurement of the moving structure.

Abstract: A method for automatic glue-spraying of stringers and inspection of glue-spraying quality based on measured data. Three-dimensional (3D) point cloud data of a stringer-skin assembly is collected by 3D laser scanner, and then processed by denoising and sampling. Feature points of an intersection line of a site to be glued of the stringer-skin assembly are extracted by K-means clustering method based on Gaussian mapping, and a minimum spanning tree is constructed based on a set of the extracted feature points. A connected region is established to obtain an initial feature intersection line of the string-skin assembly, which is optimized by random sample consensus algorithm to remove redundant small branch structures to obtain the actual glue-spraying trajectory. The quality of the glue sprayed on the stringer-skin assembly is inspected by line laser to determine positions of the defects, which are then subjected to secondary glue-spraying.