The present invention enables the structure of a neural network to be quantitatively analyzed. An analyzing unit calculates, for each of combinations of a dimension of input data and a cluster, a sum of squared errors between an output of each unit belonging to the cluster when a value of the dimension of the input data is replaced with an average value of the dimension of the input data included in learning data and an output of each unit belonging to the cluster for the input data before replacement as a relationship between the combinations, and calculates, for each of combinations of the cluster and a dimension of output data, a squared error between the value of the dimension of the output data when an output value of each unit belonging to the cluster is replaced with an average output value of each unit of the cluster when the input data included in the learning data was input and the value of the dimension of the output data before replacement as a relationship between the combinations.

A method for image processing includes obtaining a mask of a stroke from an image and identifying a plurality of cross edges for the stroke based on the mask and a reference line. The plurality of cross edges includes a group of adjacent cross edges that intersect the reference line. The method further includes (a) calculating a first vector based on positions of at least two of the cross edges in the group, (b) expanding the group, based on the first vector, to include cross edges adjacent to the group that do not intersect the reference line, (c) calculating a second vector based on positions of at least two of the cross edges in the expanded group, and (d) expanding the expanded group, based on the second vector, to include a second group of adjacent cross edges nearby the expanded group that do not intersect the reference line.

A table-image recognition device includes: an object extracting unit that extracts a plurality of objects included in a table; a set determination unit that determines whether or not every pair consisting of two objects selected from the plurality of objects is a set constituting a component specified by a column and a row of the table; a same-row determination unit that determines whether or not the objects of each pair share a same row; a same-column determination unit that determines whether or not the two objects of each pair share a same column; and a structure determining unit that determines a structure of the table by specifying the row and column to which each object belongs on the basis of the determination result.

An image processing system and method of operation includes: a source image having source pixels; homogeneous blocks in the source image having a block color; a homogeneous region in the source image formed by merging the homogeneous blocks having the block color within a color threshold; a text background region having text pixels and background pixels in the homogeneous region with the text background region having a texture feature above a texture threshold and a region size above a region size threshold; and a binary text mask representing the text pixels and the background pixels for displaying on a device.

A text recognition system and method of operation thereof including: a storage unit for storing a text unit; and a processing unit, connected to the storage unit, the processing unit including: a communication interface for receiving the text unit, a feature detection module for determining an isolated feature of the text unit, an angle detection module for determining angle features of the text unit, a feature vector module for generating a feature vector for the text unit based on the isolated feature and the angle features, and a text recognition module for determining recognized text using the feature vector for display on a display interface.

Provided are a computer program product, system, and method for training and using a vector encoder to determine vectors for sub-images of text in an image to subject to optical character recognition. A vector encoder is trained to encode images representing text into vectors in a vector space. Vectors of images representing similar text have a high degree of cohesion in the vector space. Vectors of images representing dissimilar text have a low degree of cohesion in the vector space. An input image is processed to determine sub-images of the input image that bound text represented in the input image. The sub-images are inputted to the vector encoder to output sub-image vectors. The vector encoder generates a search vector for search text. Optical character recognition is applied to at least one region of the input image including the sub-images having sub-image vectors matching the search vector.