To quickly specify a region of a character group included in an image, image obtaining means of an image processing device obtains the image data. Expansion means expands and unites a plurality of regions respectively indicating objects included in an image indicated by the image data obtained by the image obtaining means. Character region determining means determines whether all or some of the regions, which are expanded and united by the expansion means, are character regions. Processing performing means performs predetermined processing on a region that is determined by the character region determining means as a character region.

Field extraction from a form image includes identifying a target field of the form image, defining a patch from the form image based on the target field, and encoding the patch using a color encoding scheme to obtain an encoded patch. Field extraction further includes applying a trained classifier to the encoded patch to identify a relationship between a field value and a field identifier, and extracting the field value from the form image according to the relationship.

Systems and methods for automatically verifying optical character recognition (OCR) detected text of a native electronic document having an image layer comprising a matrix of pixels and a text layer comprising a sequence of characters. The method includes determining a location of OCR-detected text in the text layer of the native electronic document based on a pixel-based coordinate location of the OCR-detected text in the image layer of the native electronic document. The method also includes applying the location of the OCR-detected text to the text layer of the native electronic document to detect text in the text layer corresponding to the OCR-detected text. The method also includes rendering only the detected text in the text layer as an output when the OCR-detected text does not match the detected text in the text layer, to improve accuracy of the output text.

The present invention discloses a method, a system, an electronic device, and a medium for classifying license plates based on deep learning that are applied to an electronic device. The method includes: acquiring at least one photograph sent by a terminal device; preprocessing the acquired photograph such that the preprocessed photograph matches a plurality of input parameters of a pre-trained recognition model; and inputting the preprocessed photograph to the pre-trained recognition model to recognize corresponding vehicle use information of the photograph, and sending the corresponding vehicle use information of the photograph to the terminal device. Thus, with this disclosure, the use of a vehicle in a photograph can be automatically and accurately recognized and further the photographs can be accurately classified, thereby improving the accuracy as well as the efficiency.

A method of detecting background pixels in a video, the video including a sequence of frames, each frame having a pitch and people. The method includes segmenting the pitch and determining a height of one of the people. An inclusion mask for the sequence of frames is then created using the pitch segmentation and the people's height. Each of the sequence of frames is segmented into foreground and background pixels. The method then creates a foreground appearance model from the foreground pixels inside said inclusion mask over the sequence of frames, and a background appearance model from the background pixels inside said inclusion mask over the sequence of frames. The method then uses the created foreground and background appearance models to segment one of the sequence of frames into definite foreground, definite background and uncertain pixels.

A computer implemented a method and system for enrichment of OCR extracted data is disclosed comprising of accepting a set of extraction criteria and a set of configuration parameters by a data extraction engine. The data extraction engine captures data satisfying an extraction criteria using the configuration parameters and adapts the captured data using a set of domain specific rules and a set of OCR error patterns. A learning engine generates learning data models using the adapted data and the configuration parameters and the system dynamically updates the extraction criteria using the generated learning data models. The extraction criteria comprise one or more extraction templates wherein an extraction template includes one of a regular expression, geometric markers, anchor text markers and a combination thereof.

Disclosed are systems and methods for automatically verifying text of a native digital document having an image layer and a text layer. The text is detected by optical character recognition (OCR) of the image layer, and is compared to text at a corresponding location in the text layer. Normalization processing is performed on both the detected image-layer text and the text-layer text. When the image-layer text and the text-layer text do not match, the text-layer text may be used or an icon indicating that the image-layer text and the text-layer text do not match is rendered and displayed.

A method for recognizing characters in an image of a document having at least one alphanumeric field. The method includes the steps of enhancing an image contrast to highlight the characters in the image; detecting contours of objects in the image to create a mask that highlights the characters; segmenting the image using a tree with connected components and applying the mask thereto in order to extract the characters from the image; performing a character recognition on the extracted objects. A device for implementing the method.

A method of generating handwriting information about handwriting of a user includes determining a first writing focus and a second writing focus; sequentially shooting a first local writing area, which is within a predetermined range from a first writing focus, and a second local writing area, which is within a predetermined range from a second writing focus; obtaining first handwriting from the first local writing area and second handwriting from the second local writing area; combining the first handwriting with the second handwriting; and generating the handwriting information, based on a result of the combining.

Disclosed are systems and methods for automatically verifying text of a native digital document having an image layer and a text layer. The text is detected by optical character recognition (OCR) of the image layer, and is compared to text at a corresponding location in the text layer. Normalization processing is performed on both the detected image-layer text and the text-layer text. When the image-layer text and the text-layer text do not match, the text-layer text may be used or an icon indicating that the image-layer text and the text-layer text do not match is rendered and displayed.