Sap Se

Marco Spinaci

Johannes Hoehne

Anoop Raveendra Katti

Christian Reisswig

Disclosed herein are system, method, and computer program product embodiments for optical character recognition (OCR) pre-processing using machine learning. In an embodiment, a neural network may be trained to identify a standardized document rotation and scale expected by an OCR service performing character recognition. The neural network may then analyze a received document image to identify a corresponding rotation and scale of the document image relative to the expected standardized values. In response to this identification, the document image may be modified in the inverse to standardize the rotation and scale of the document image to match the format expected by the OCR service. In some embodiments, a neural network may perform the standardization as well as the character recognition using a shared computation graph.

Rotation and scaling for optical character recognition using end-to-end deep learning

In some embodiments, a method inputs a set of images into a network and trains the network based on a classification of the set of images to one or more characters in a set of characters. The method obtains a set of encodings for the one or more characters based on a layer of the network that restricts the output of the layer to a number of values. Then, the method stores the set of encodings for the one or more characters, wherein an encoding in the set of encodings is retrievable when a corresponding character is determined.

Visually-aware encodings for characters

Disclosed herein are system, method, and computer program product embodiments for providing object detection and filtering operations. An embodiment operates by receiving an image comprising a plurality of pixels and pixel information for each pixel. The pixel information indicates a bounding box corresponding to an object within the image associated with a respective pixel and a confidence score associated with the bounding box for the respective pixel. Pixels that do not correspond to a center of at least one of the bounding boxes are iteratively removed from the plurality of pixels until a subset of pixels each of which correspond to a center of at least one of the bounding boxes remains. Based on the subset, a final bounding box associated with each object of the image is determined based on an overlapping of the bounding boxes of the subset of pixels and the corresponding confidence scores.

Object detection and candidate filtering system

Disclosed herein are system, method, and computer program product embodiments for optical character recognition using end-to-end deep learning. In an embodiment, an optical character recognition system may train a neural network to identify characters of pixel images, assign index values to the characters, and recognize different formatting of the characters, such as distinguishing between handwritten and typewritten characters. The neural network may also be trained to identify, groups of characters and to generate bounding boxes to group these characters. The optical character recognition system may then analyze documents to identify character information based on the pixel data and produce segmentation masks, such as a type grid segmentation mask, and one or more bounding box masks. The optical character recognition system may supply these masks as an output or may combine the masks to generate a version of the received document having optically recognized characters.

Growing community of inventors

Berlin, Germany

Marco Spinaci