Generation of image corresponding to input text using multi-text guided image cropping

    公开(公告)号:US12131406B2

    公开(公告)日:2024-10-29

    申请号:US18052870

    申请日:2022-11-04

    申请人: Lemon Inc.

    IPC分类号: G06T11/00 G06F40/40 G06T5/70

    CPC分类号: G06T11/00 G06F40/40 G06T5/70

    摘要: Systems and methods are provided that include a processor executing a program to receive an input from a user, where the input including a first input text and a second input text. The processor is further configured to provide an initial image and, for a predetermined number of iterations, define a first and second regions of the initial image associated with the first and second input texts, respectively, define a plurality of patches of the initial image, input the initial image into a diffusion process to generate a processed image, back-propagate the processed image through a text-image match gradient calculator by generating an image embedding based on the processed image, generating a text embedding based on the region and the input text that are associated with a patch, and calculating a differential between the image embedding and the text embedding.

    Accelerating data processing by offloading thread computation

    公开(公告)号:US12118397B2

    公开(公告)日:2024-10-15

    申请号:US17945843

    申请日:2022-09-15

    IPC分类号: G06F9/50 G06F9/48 G06F15/78

    摘要: The present disclosure describes techniques for accelerating data processing by offloading thread computation. An application may be started based on creating and executing a process by a host, the process associated with a plurality of threads. Creating a plurality of computation threads on a storage device may be requested based on determining that the storage device represents a computational storage. The plurality of computation threads may be created based on preloading a plurality of libraries in the storage device. The plurality of libraries may comprise executable codes associated with the plurality of threads. Data processing associated with the plurality of threads may be offloaded to the storage device using the plurality of computation threads. Activities associated with the plurality of computation threads may be managed by the process.

    SECURE COMPUTATION AND COMMUNICATION
    6.
    发明公开

    公开(公告)号:US20240340178A1

    公开(公告)日:2024-10-10

    申请号:US18297339

    申请日:2023-04-07

    申请人: Lemon Inc.

    摘要: Methods and systems for secure computation and communication are provided. The method includes transforming identifications of a first dataset using a first transforming scheme, and transforming attributes of the first dataset using a second transforming scheme. The method also includes dispatching the transformed first dataset, receiving a second dataset, transforming identifications of the received second dataset, dispatching the identifications of the transformed received second dataset, and receiving a set of identifications. The method further includes generating a first intersection of the received set of identifications and the transformed received second dataset, generating a first share based on the first intersection, receiving a second share, and constructing a result based on the first share and the second share.

    METHOD AND APPARATUS FOR KNOWLEDGE GRAPH CONSTRUCTION, STORAGE MEDIUM, AND ELECTRONIC DEVICE

    公开(公告)号:US20240330373A1

    公开(公告)日:2024-10-03

    申请号:US18573944

    申请日:2022-08-12

    申请人: Lemon Inc.

    IPC分类号: G06F16/903

    CPC分类号: G06F16/90344

    摘要: The disclosure relates to a method and apparatus for knowledge graph construction, a storage medium, and an electronic device. The method comprises: obtaining a target entity identifier and determining an industry type label corresponding to the target entity identifier; determining a target industry attribute table based on a predetermined correspondence among the industry type label, an industry type, and an industry attribute table; obtaining target attribute values of the target entity identifier from a public database based on respective target attribute names in the target industry attribute table, to obtain a target attribute of the target entity identifier, wherein the target attribute characterizes a key-value pair consisting of the target attribute name and the target attribute value; and constructing a knowledge graph based on an entity characterized by the target entity identifier, the industry type label, and the target attribute.

    METHOD FOR MANUFACTURING OPTICAL WAVEGUIDE AND OPTICAL WAVEGUIDE

    公开(公告)号:US20240329311A1

    公开(公告)日:2024-10-03

    申请号:US18617623

    申请日:2024-03-26

    IPC分类号: G02B6/136 G02B6/124

    CPC分类号: G02B6/136 G02B6/124

    摘要: The embodiments of the present disclosure provide a method for manufacturing an optical waveguide and an optical waveguide, the method for manufacturing includes: providing a substrate; forming a first thin film layer, a second thin film layer and a sacrificial layer on the substrate in a stacked manner, a refractive index of the first thin film layer is larger than 2; exposing and developing the sacrificial layer so that the sacrificial layer forms a first mask layer; etching the second thin film layer by taking the first mask layer as mask so that the second thin film layer forms a second mask layer; removing the first mask layer, and etching the first thin film layer by taking the second mask layer as mask so that the first thin film layer forms a grating layer; and removing the second mask layer to form the optical waveguide comprising the grating layer and the substrate.

    Coding of motion information
    10.
    发明授权

    公开(公告)号:US12101468B2

    公开(公告)日:2024-09-24

    申请号:US17933755

    申请日:2022-09-20

    申请人: Lemon Inc.

    摘要: Implementations of the present disclosure provide a solution for reordering merge candidates. In this solution, during a conversion between a current video block of a video and a bitstream of the video, a template matching cost associated with each of a plurality of candidates of motion information for the current video block is determined. The template matching cost is determined based on a first difference between a set of neighboring samples of the current video block and a set of reference samples corresponding to the set of neighboring samples, and a second difference between a set of reconstructed samples neighboring to the set of neighboring samples and the set of reference samples. The set of reference samples is determined based on motion information indicated by a respective candidate. Then, the plurality of candidates of motion information is reordered based on the template matching cost.