7results about How to "Shorten calculation time" patented technology

The present application relates to the technical field of three-dimensional reconstruction, and in particular to an object three-dimensional reconstruction system based on deep learning, which introduces an adaptive cost aggregation method with visibility perception for cost volume aggregation, acquires the visibility of pixel points in the view through a network, and can improve the reconstruction integrity of the occluded area; a variance-based method is used to predict the disparity range of each pixel, and a spatially-varying depth hypothesis surface is constructed for depth estimation in the next stage, and a residual and channel attention guided fusion depth map optimization module is proposed in the last stage to obtain an optimized depth map; an improved depth map fusion algorithm is used to combine the pixel point and 3D point re-projection error for consistency checking to obtain a dense point cloud. Quantitative and qualitative comparison results of the present application and other methods on the DTU dataset show that the present application can reconstruct a scene with better details, and achieves the purposes of reducing GPU memory consumption and computation time.

The present application relates to the technical field of image recognition, and discloses a detection method, device and equipment for distracted driving behavior and a storage medium, which are used to improve the recognition accuracy of distracted driving behavior. The detection method for distracted driving behavior comprises the following steps: obtaining initial driving video data of a driver, performing data processing on the initial driving video data, and obtaining to-be-recognized behavior video data; calling a preset distracted driving behavior model to recognize the to-be-recognized behavior video data; if the driver has candidate distracted driving behavior, performing special behavior detection on the driver and vehicle driving condition detection on a target vehicle in which the driver is located; and if the driver has no special behavior and the target vehicle is in a normal vehicle driving condition, it is determined that the driver has target distracted driving behavior.

The application provides a kind of three-dimensional fracture whole process simulation method of multiphase medium material mesostructure based on space-time sequence prediction, comprising the following steps: S1, sample data acquisition;S2, deep learning model network building and training;S3, model verification: the historical surface crack information of composite material is input into the trained generation network, and the subsequent fracture damage process of composite material can be obtained, the morphology between fracture surfaces is quantified and compared by using Wasserstein distance, and the performance of the model is judged.The application can predict the fracture surface inside the entity according to the historical apparent crack of the composite material without the component distribution information of the composite material;The network model is trained completely from the perspective of observation data, and theoretically, the prediction result can be almost consistent with the observed result in the experiment;The calculation efficiency is high, and the calculation time can be shortened from weeks to seconds;Pixel-level modeling can be realized;The fracture process inside the material can be predicted according to the apparent crack of the material.

The invention provides a replenishment configuration method and device, electronic equipment, a storage medium and a computer program product, and relates to the technical field of supply chains. The method comprises the following steps: according to an execution sequence of a plurality of workflow definitions in a preset workflow arrangement, sequentially reading configuration data corresponding to each workflow definition based on a spark calculation engine for processing, and determining replenishment information corresponding to the preset workflow arrangement; wherein the execution of each workflow definition needs a processing result corresponding to the previous workflow definition. Through the scheme in the embodiment of the invention, the efficiency of replenishment calculation can be improved.

The specification discloses a navigation structure generation method and device, a storage medium and an electronic device, wherein the method comprises: obtaining three-dimensional modeling data corresponding to each building in a target geographical area, and a regional image corresponding to the target geographical area; determining a building road network of at least one floor in each building based on the three-dimensional modeling data; generating a road network structure of the target geographical area based on an outdoor road network in the regional image and the building road network; the outdoor road network is other road network in the regional image except the building road network; mapping the road network structure into a planar graph data structure; obtaining a structure parameter corresponding to the graph data structure; and using the specification, the three-dimensional data of the target geographical area is converted into planar data, and the structure parameter of the graph data structure corresponding to the target geographical area is calculated according to the planar data, so as to improve the accuracy and efficiency of the navigation path.

The application provides a privacy-protected longitudinal federated boosting tree method and relates to the technical field of longitudinal federated learning; the method comprises the following steps: combining secret sharing binary matrix and gradient information to obtain a shared matrix and secret sharing; each training terminal performs row permutation on the shared matrix based on a permutation triple and secret shares the permutation result; linear gain scores of candidate splits are calculated and secret shared, and an indication vector corresponding to the maximum linear gain score is calculated; according to the number of candidate splits, corresponding columns are selected from the matrix, and the indication vector is used to calculate the sample and the permutation matrix of the child node; the splitting step is recursively executed to complete the splitting from the root node to the leaf node, the weight of the leaf node is calculated and stored in a preset set. The binary permutation matrix of the root node of the next decision model is updated based on the weight of the leaf node, and the calculation is continued until all decision tree models are generated; the method can solve the problem of low efficiency of the existing longitudinal federated gradient boosting tree protocol.

The invention relates to the field of vacuum system design and calculation, and particularly discloses a vacuum system topology analysis and evacuation time calculation method based on a dual-accounting method, which comprises a main core algorithm, establishment of a directed graph based on connection points of vacuum components, discovery of all physical flow paths from a source point to a sink point by adopting depth-first search, and calculation of the vacuum system topology analysis and evacuation time. The three types of structure fragments are introduced into a virtual component for recursive simplification to obtain equivalent conductance; according to the auxiliary core algorithm, key contacts are recognized, the inflow and outflow relation of the key contacts is judged, serial-parallel equivalent result write-back is locally completed, rapid estimation of the global equivalent through-current capacity is given, and the main core algorithm is verified; and the calculation efficiency of the vacuum system is improved by adopting three-stage pre-calculation. According to the method, non-uniqueness and misrecognition caused by manual splitting are reduced through a dual-core cooperation mechanism combining a main core algorithm and an auxiliary core algorithm, and the series-parallel misjudgment rate of a complex network is reduced.