An image processing apparatus includes circuitry to determine a type of a document read by a scanner, set a top-bottom determination method based on the type of the document, and determine a top-bottom orientation of a target image by the top-bottom determination method. The target image is obtained by reading the document with the scanner.

Provided is an image reading system that divides, with respect to image data obtained by performing a double sided reading in a state where a booklet is opened, cover sheet image data into two parts corresponding to a pair of cover sheets to arrange the two parts at a front and an end, respectively, and arranges main text image data between the front and the end, and then generates an image file from each of the arranged image data.

A method involves detecting primary entities in a document, involving determining that a subset of the primary entities are associated with a first primary entity type, and determining a second primary entity type of one of the primary entities. The method further involves processing the primary entity of the second primary entity type to determine a secondary entity type of the primary entity. The secondary entity type is a subcategory of the second primary entity type. The method also involves hierarchically organizing the primary entities into a document layout structure that includes a top level and a child level. The top level is established by the first subset of primary entities based on the first primary entity type identifying the first subset as headings, and the child level is established by the primary entity based on the second primary entity type, the child level identifying the secondary entity type.

An image processing apparatus includes a layout analyzing part that executes layout analysis for image data, an extraction part that extracts a diagrammatic representation from the image data by using a result of the layout analysis, a character recognizing part that executes character recognition for a partial area having a high probability of presence of a character string in a relationship with the extracted diagrammatic representation, and an erecting direction deciding part that decides an erecting direction of the image data by using a result of the character recognition.

An apparatus determines whether a reflecting portion that strongly reflects a light of a light source is included in a captured image. And a message for prompting the user to change the shooting method is displayed when it is determined that the reflecting portion is included in the captured image.

An image forming apparatus, a scanned image correction method of an image forming apparatus, and a non-transitory computer-readable recording medium are provided. The image forming apparatus includes a scan unit to scan a document to generate a scanned image and a processor to detect a skew angle of the scanned image, determine a reference point on the basis of a position of a content in the scanned image, and rotate the scanned image around the determined reference point to correct the skew angle.

An embodiment text recognition device includes a character position recognizer configured to recognize individual characters in an image, and the character position recognizer is also configured to recognize a position of each of the individual characters, a correction processor configured to set a main region, the correction processor further being configured to perform one or both of correcting a slope of the main region and magnification calibration for at least one character recognized by the character position recognition part, and a text recognizer configured to perform text recognition in the main region corrected by the correction processor.

In some embodiments, apparatuses and methods are provided herein useful to detecting text of interest. In some embodiments, there is provided a system to detect vertically oriented text of interest including at least one camera and a control circuit configured to execute a trained machine learning model to automatically detect vertically oriented text of interest on an object of interest. The trained machine learning model is at least trained on a first data set including a plurality of captured digital images each depicting the object of interest, and a second data set including a plurality of augmented digital images each depicting a captured digital image augmented with a synthetic text image including randomly generated text on a randomly selected background image.

The present disclosure discloses a character offset detection method and system. The method includes: acquiring a text image; performing character separation based on the text image to obtain a character text region; calculating a center point of each rectangular box in the character text region to obtain a center point set; determining an optimal fitted curve based on the center point set; and analyzing character offset based on the optimal fitted curve to obtain an offset result. The present disclosure realizes detection of the character offset based on curve fitting, so that the accuracy of detection is improved.

A method and apparatus for locating dot text in an image are described. A set of dots is extracted. A determination of whether a first region of interest (ROI) including the set of dots satisfies selection criteria is performed, where the first region of interest is oriented based on results from a principal component analysis of the set of dots. Responsive to determining that the first ROI does not satisfy the selection criteria, performing the following: removing an outlier dot from the first set of dots to obtain a second set of dots; when the second ROI satisfies the selection criteria, outputting the second ROI as a location of the dot text in the image, and when the second region of interest does not satisfy the selection criteria, repeating the operations until a resulting ROI is determined to satisfy the selection criteria.