Abstract: Disclosed is a graph convolution auto-encoder based method for computing road network similarity. The method includes: creating a dual graph of a road network, and giving road network space feature information to nodes of the dual graph from three aspects of global, local and connection characteristics on the basis of a relation principle between an entire structure and parts of the structure, such that a quantitative expression of a road network graph structure is obtained; aggregating and updating node feature information and structure information of a road network graph by the graph convolution auto-encoder, and forming a deep understanding of the road network, such that a coded expression of node information of the road network is obtained; and mapping a complex high-dimensional feature space to an easy-to-measure low-dimensional feature space through an average pooling operation, so as to obtain a set of feature vectors, and computing the similarity.

Abstract: Disclosed in the present disclosure is a similarity calculation method considering multiple features for a multi-scale road network, which includes structural similarity calculation of skeleton lines and local feature similarity calculation of the multi-scale road network. A road network stroke is generated, skeleton lines of the road network are extracted and transformed into a structure tree, and the skeleton similarity of the road network is evaluated by calculating structural similarity of the structure tree of the multi-scale road network. Topological similarity of the multi-scale road network is calculated by using a difference matrix of a conceptual domain graph of road meshes, geometric similarity of the multi-scale road network is calculated by using a density of road meshes, and local similarity of the multi-scale road network is obtained by integrating the hierarchy of the road network into the topological and geometric similarity calculation in the form of matrices.

Abstract: Disclosed in the present disclosure is a commutative encryption and watermarking method based on a chaotic system and a zero watermark for vector geospatial data. According to the method, firstly, the vector geospatial data are scrambled and encrypted by using chaotic sequences generated by a composite chaotic system. Then, vector geospatial elements are randomly combined in pairs. A feature matrix is constructed according to the number of vertex coordinates of the vector geospatial elements in combinations, and the parity of the number. Finally, an XOR operation is performed on the feature matrix and the watermark image to construct a zero watermark image, and the zero watermark is constructed through invariant features of the vector geospatial data.

Abstract: Disclosed in the present disclosure is a classification method based on skeleton lines for building shapes. The method includes: (1) expanding a template library of buildings by combining building shape classification in architecture and building shape features in real life on the basis of inheriting advantages of a traditional template matching method; (2) avoiding influence of small depressions and protrusions of the buildings on overall shapes by constructing a least-squares template; (3) extracting the skeleton lines of the buildings and calculating feature vectors of the buildings; and (4) calculating similarity between feature vectors of skeleton lines of the buildings and skeleton lines of templates by using cosine similarity, and selecting the template with the highest similarity as a classification result of the building shapes.

Abstract: A scalable neural network accelerator may include a first circuit for selecting a sub array of an array of registers, wherein the sub array comprises LH rows of registers and LW columns of registers, and wherein LH and RH are integers. The accelerator may also include a register for storing a value that determines LH. In addition, the accelerator may include a first load circuit for loading data received from the memory bus into registers of the sub array.

Abstract: The disclosure provides a reversible watermarking method for the oblique photography three-dimensional models with controllable accuracy. According to the disclosure: firstly, by using the global stability of the average included angle of vertex normal vector, the feature points of the oblique photography three-dimensional model are extracted; secondly, the mapping relationship is established by the ratio of the distance between feature points and non-feature points, and the vertexes are grouped, and each group consists of a feature point corresponding to several non-feature points; finally, the spherical coordinate system is constructed by taking the feature points as the coordinate origin in grouping, and the watermark is embedded by modifying the radius of the coordinate system.

Abstract: A scalable neural network accelerator may include a first circuit for selecting a sub array of an array of registers, wherein the sub array comprises LH rows of registers and LW columns of registers, and wherein LH and RH are integers. The accelerator may also include a register for storing a value that determines LH. In addition, the accelerator may include a first load circuit for loading data received from the memory bus into registers of the sub array.

Abstract: The present disclosure discloses a method and an apparatus for processing seismic data, and belongs to the field of geological surveys. The method comprises: stacking seismic trace gathers in a predetermined range among S seismic trace gathers after a Normal Move Out (NMO) correction processing to obtain a model trace, S being an integer; calculating a correlation coefficient of each seismic trace gather with the model trace, and selecting a K-th seismic trace gather with a maximum correlation coefficient; calculating an optimum point of each seismic trace gather from the K-th seismic trace gather to two sides orderly; and performing a residual NMO correction of the seismic trace gathers according to the optimum points.

Abstract: The present disclosure discloses a method and an apparatus for processing seismic data, and belongs to the field of geological surveys. The method comprises: stacking seismic trace gathers in a predetermined range among S seismic trace gathers after a Normal Move Out (NMO) correction processing to obtain a model trace, S being an integer; calculating a correlation coefficient of each seismic trace gather with the model trace, and selecting a K-th seismic trace gather with a maximum correlation coefficient; calculating an optimum point of each seismic trace gather from the K-th seismic trace gather to two sides orderly; and performing a residual NMO correction of the seismic trace gathers according to the optimum points.

Abstract: In one embodiment of a header-compression method, a 32-bit timestamp value is divided by a 16- or 8-bit stride value using a plurality of 16/8-bit division operations, each performed using a corresponding hardware instruction issued to an arithmetic logic unit (ALU) of the corresponding communication device, such as an access terminal or a base station of a communication system. When specialized 32/16-bit and/or 32/8-bit division-logic circuitry is not available in the ALU, embodiments of the header-compression method can advantageously be used to improve the speed and efficiency of timestamp compression in communication devices.

Abstract: A seamless power transfer apparatus comprising a solar grid and a utility power grid. The apparatus comprises a smart switch electrically connected to the solar grid and to the utility grid. The smart switch does not allow the solar grid to supply power to the utility power grid at any time.

Abstract: In one embodiment of a header-compression method, a timestamp value is divided by a stride value using a plurality of binary-shift operations corresponding to a Taylor expansion series of the reciprocal stride value in a base of ½. When the division-logic circuitry of an arithmetic logic unit in the corresponding communication device is not designed to handle operands that can accommodate the length of the timestamp and/or stride values, the header-compression method can advantageously be used to improve the speed and efficiency of timestamp compression in communication devices.

Abstract: In one embodiment of a header-compression method, a 32-bit timestamp value is divided by a 16- or 8-bit stride value using a plurality of 16/8-bit division operations, each performed using a corresponding hardware instruction issued to an arithmetic logic unit (ALU) of the corresponding communication device, such as an access terminal or a base station of a communication system. When specialized 32/16-bit and/or 32/8-bit division-logic circuitry is not available in the ALU, embodiments of the header-compression method can advantageously be used to improve the speed and efficiency of timestamp compression in communication devices.

Abstract: In one embodiment of a header-compression method, a timestamp value is divided by a stride value using a plurality of binary-shift operations corresponding to a Taylor expansion series of the reciprocal stride value in a base of ½. When the division-logic circuitry of an arithmetic logic unit in the corresponding communication device is not designed to handle operands that can accommodate the length of the timestamp and/or stride values, the header-compression method can advantageously be used to improve the speed and efficiency of timestamp compression in communication devices.

Abstract: In one embodiment, a method includes receiving a media signal associated with a session established based on a session request. The session request is defined at a session exchange device associated with a media over internet protocol (MoIP) network. The method also includes sending a portion of the media signal to a dual-tone multi-frequency (DTMF) analysis module during a sampling time period after a non-sampling time period.

Abstract: In one embodiment, a method includes receiving a first internet protocol (IP) packet having information associated with a dual-tone multi-frequency (DTMF) signal. The information of the first IP packet is configured based on a protocol associated with a first layer of a media over internet protocol (MoIP) network. The first IP packet is associated with a destination endpoint. The method also includes producing a second IP packet having information associated with the DTMF signal. The information of the second IP packet is configured based on a protocol associated with a second layer of the MoIP network and a DTMF conversion policy associated with the destination endpoint. The second layer is different than the first layer.

Abstract: In one embodiment, a method includes receiving a media signal associated with a session established based on a session request. The session request is defined at a session exchange device associated with a media over internet protocol (MoIP) network. The method also includes sending a portion of the media signal to a dual-tone multi-frequency (DTMF) analysis module during a sampling time period after a non-sampling time period.

Abstract: In one embodiment, a method includes receiving a first internet protocol (IP) packet having information associated with a dual-tone multi-frequency (DTMF) signal. The information of the first IP packet is configured based on a protocol associated with a first layer of a media over internet protocol (MoIP) network. The first IP packet is associated with a destination endpoint. The method also includes producing a second IP packet having information associated with the DTMF signal. The information of the second IP packet is configured based on a protocol associated with a second layer of the MoIP network and a DTMF conversion policy associated with the destination endpoint. The second layer is different than the first layer.