A method and system for training a neural network are provided. The method includes receiving an input image, selecting at least one data augmentation method from a pool of data augmentation methods, generating an augmented image by applying the selected at least one data augmentation method to the input image, and generating a mixed image from the input image and the augmented image.

Systems and methods are provided to perform context-aware text extraction.

A technique for document management includes performing a text-based analysis of a document submitted for signature and/or storage. The text-based analysis searches for terms and/or other content that indicate a future action to be performed and a corresponding action date by which to perform the action. In response to detecting such terms and/or other content, the technique further includes creating metadata indicative of the action and associated action date and storing such metadata in a database. The technique further includes querying the database, e.g., prior to or upon the action date, and creating a modified version of the document to implement the detected action.

Certain aspects of the present disclosure provide techniques for extracting data from a document. An example method generally includes identifying a bounding polygon of the region from an electronic image of the document and extracting data from within the bounding polygon of the region. The method further includes generating revised extracted data based on the extracted data, and combining the revised extracted data with other data extracted from the electronic image of the document to generate input data for a data processing application.

The invention relates to a neural network for semantic segmentation of a document with complex text. The network comprises a first multilayer neural encoding chain of an initial image file of the document to be processed, a second multilayer neural encoding chain of a mask image file of the document to be processed, a multilayer neural decoding chain connected to the outputs of the first and second encoding chains, a first bridge of parallel residual connections between the layers of the first encoding chain and the layers of the decoding chain, a second bridge of parallel residual connections between the layers of the second encoding chain and the layers of the decoding chain, the residual connections generating files of the same size that are connected after a layer of the neural decoding chain generating a file of the same size.

A method for training an image editing model includes steps described below. Covering processing is performed on a region of interest determined in an original image so that a background image sample is formed, and content corresponding to the region of interest is determined as a sample of content of interest; the background image sample and the sample of the content of interest are input into an image editing model; fusion processing is performed on a background image feature and a feature of the region of interest by using the image editing model so that a fusion feature is formed; an image reconstruction operation is performed according to the fusion feature by using the image editing model so that a reconstructed image is output; and optimization training is performed on the image editing model according to a loss relationship between the reconstructed image and the original image.

The present disclosure is directed to systems and methods for analyzing digital images to determine alphanumeric strings depicted in the digital images. An electronic device may generate a set of filtered images using a received digital image. The electronic device may also perform an optical character recognition (OCR) technique on the set of filtered images, and may filter out any of the set of filtered images according to a set of rules. The electronic device may further identify a set of common elements representative of the alphanumeric string depicted in the digital image, and determine a machine-encoded alphanumeric string based on the set of common elements.

This application discloses a data processing method and apparatus, a computer device, and a non-transitory computer-readable storage medium in the technical field of computers. This application, for textual data and picture data of an article, extracts a textual feature and a picture feature, respectively, and predicts an article classification to which the article belongs using a cross-modal interaction feature between the textual feature and picture feature. At the same time, this application considers the contribution degree of each of a textual modality and a picture modality to the article classification, rather than determining from a textual perspective only. In addition, the extracted cross-modal interaction feature is not a simple concatenation of the textual feature and the picture feature, which can reflect richer and deeper inter-modal interaction information, and greatly improve the identification accuracy of the article classification. Furthermore, it can improve the discovering accuracy of high-quality articles in the scene of identifying high-quality articles.

The present disclosure is directed to systems and methods for analyzing digital images to determine alphanumeric strings depicted in the digital images. An electronic device may generate a set of filtered images using a received digital image. The electronic device may also perform an optical character recognition (OCR) technique on the set of filtered images, and may filter out any of the set of filtered images according to a set of rules. The electronic device may further identify a set of common elements representative of the alphanumeric string depicted in the digital image, and determine a machine-encoded alphanumeric string based on the set of common elements

The present disclosure is directed to systems and methods for analyzing digital images to determine alphanumeric strings depicted in the digital images. An electronic device may generate a set of filtered images using a received digital image. The electronic device may also perform an optical character recognition (OCR) technique on the set of filtered images, and may filter out any of the set of filtered images according to a set of rules. The electronic device may further identify a set of common elements representative of the alphanumeric string depicted in the digital image, and determine a machine-encoded alphanumeric string based on the set of common elements.