3 results about "Local selection" patented technology

The application discloses a crop disease identification method based on local selection and feature interaction, which adopts a deep learning image recognition technology to identify and classify crop disease images. The application firstly performs data preprocessing on the disease images, and then extracts leaf disease features through an LSFI-ViT model constructed by the application. The LSFI-ViT takes a ViT as a backbone network to extract global features of the whole image, a local selection module LSM and a local-global feature interaction attention encoder LGFIA-Former are connected in series to form an extended branch, the LSM is responsible for selecting a local area with distinguishing features in the image, and the LGFIA-Former simultaneously processes local and global feature sequence inputs through the LGFIA to obtain interactive features of the local-global. Two groups of features are respectively sent to different fully connected layers to obtain two groups of prediction results, and finally, effective identification of different crop diseases and severity is realized in an integrated voting manner. The application effectively solves the problem of unsatisfactory crop disease and severity identification.

The application relates to the technical field of graph data processing, and discloses a graph alignment method and device based on GW distance, a system and a storage medium. The method first constructs an initial coupling matrix and obtains an optimization matrix for graph alignment from multiple rounds of global updating operations, and the construction and updating operations of the coupling matrix are based on a related algorithm of GW distance. Secondly, in the process of each round of global updating, a local selection block is first confirmed to ensure that an adaptive dynamic region selection strategy can be adopted in stages. Finally, the iterative updating operation in the current round of global updating is limited within the local selection block, and updating is performed according to a coupling submatrix obtained on the coupling matrix according to the local selection block, to obtain an updating submatrix, and then the coupling matrix is updated in a local range according to the updating submatrix, so that the updating range can be compressed to the local selection block range of each round under the condition that the total round number of direct updating is the same, thereby reducing the calculation complexity of the GW distance in the graph alignment task.

The invention relates to the technical field of printed circuit board manufacturing, discloses a dynamic etching compensation method and system based on an Incam Pro platform, and realizes intelligent compensation processing of objects such as wiring, bonding pads, copper sheets and the like by integrating attribute recognition, automatic parameter loading, multi-stage dynamic compensation and local selection functions in Incam Pro software. The system can automatically obtain the newest process rule from the DPM database, supports compensation sequence custom and result visual verification, and remarkably improves compensation precision, automation level and process consistency. According to the method, an existing CAM platform does not need to be replaced, calling is carried out in a script mode, and the method has the advantages of being convenient to deploy, high in compatibility and high in practical value and is suitable for HDI, IC carrier boards, high-speed communication boards and other high-end PCB manufacturing scenes.