Method and apparatus for character recognition

A method for forming a text file of characters in pixel image data comprising the steps of:

• inputting pixel image data, wherein said inputting step includes the step of converting the pixel image data into binary pixel image data in the event that the pixel image data is not binary pixel image data;
• selecting blocks of pixel image data, said selecting step including, in order, the steps of outlining contours of connected components in the pixel image data, determining whether the outlined connected components include text units or non-text units based on the size of the outlined connected components, selectively connecting text units widthwisely to form text lines based on proximity of adjacent text units, and vertically connecting a plurality of the formed text lines to form text blocks based on proximity of adjacent text lines and on the position of non-text units between text lines, at least one formed text block having a plurality of the formed text lines;
• forming a hierarchical tree based on the connected components outlined in said outlining step and designating as a descendent a connected component which is within a rectangle formed around another connected component;
• modifying the hierarchical tree based on the connecting steps;
• segmenting text blocks into lines of pixel image data;
• cutting characters from the line segmented in said segmenting step;
• recognizing characters cut in said cutting step and deriving character codes based on said recognition;
• and
• storing the character codes in a computer text file in accordance with an order established by the modified hierarchical tree.

An apparatus for recognizing characters in pixel image data, comprising:

• selecting means for selecting blocks of pixels from the pixel image data, said selecting means including outlining means for outlining contours of connected components in the pixel image data, determining means for determining whether the outlined connected components include text units or non-text units, tree forming means for forming a hierarchical tree based on the outlined connected components, first connecting means for selectively connecting text units widthwisely to form text lines, second connecting means for vertically connecting the formed text lines to form text blocks, at least one formed text block having a plurality of the formed text lines, and modifying means for modifying the hierarchical tree based on the connections performed by said first and second connecting means, wherein said tree forming means includes means for designating as a descendent a connected component which is within a rectangle formed around another connected component;
• segmenting means for segmenting text blocks into lines of pixel image data;
• cutting means for cutting characters from the line segmented by said segmenting means;
• recognizing means for recognizing characters cut by said cutting means;
• and
• storing means for storing characters recognized by said recognizing means in accordance with an order established by said block selecting means.

An apparatus for forming a text file of characters in pixel image data comprising:

• input means for inputting pixel image data, said input means including converting means for converting the pixel image data into binary pixel image data in the event that the pixel image data is not binary pixel data;
• selecting means for selecting blocks of pixel image data, said selecting means including outlining means for outlining contours of connected components in the pixel image data, determining means for determining whether the outlined connected components include text units or non-text units based on the size of the outlined connected components, first connecting means for selectively text units widthwisely to form text lines based on proximity of adjacent text units, and second connecting means for vertically connecting the formed text lines to form text blocks based on proximity of adjacent text lines and on the position of non-text units between text lines, at least one formed text block including a plurality of the formed text lines;
• forming means for forming a hierarchical tree based on the connected components outlined by said outlining means, and for designating as a descendent a connected component which is within a rectangle formed around another connected component;
• modifying means for modifying the hierarchical tree based on the connections performed by said first and second connecting means;
• segmenting means for segmenting text blocks into lines of pixel image data;
• cutting means for cutting characters from the line segmented by said segmenting means;
• recognizing means for recognizing characters cut by said cutting means and for deriving character codes based on said recognition;
• and
• storing means for storing the character codes in a computer text file in accordance with an order established by the modified hierarchical tree.

A method for recognizing characters in pixel image data comprising the steps of:

• selecting blocks of pixels from the pixel image data, said selecting step including the steps of outlining connected components in the pixel image data, forming a hierarchical tree based on the outlined connected components, connecting the outlined connected components to form text lines and text blocks, and modifying the hierarchical tree based on the connected and outlined connected components, wherein said forming step comprises the step of designating as a descendent a connected component which is within a rectangle formed around another connected component;
• segmenting text blocks into lines of pixel data by adaptively dividing the text blocks into at least one column based on a horizontal projection of pixel density across said at least one column;
• cutting characters from the line segmented in said segmenting step, said cutting step including a first cutting layer step in which non-touching and non-overlapping characters are cut from the line segment and at least one additional cutting layer step in which touching characters are cut, said additional cutting layer step being carried out in accordance with whether or not the line characters are monospaced;
• recognizing the characters cut in said cutting step;
• and
• storing characters recognized in said recognizing step in accordance with an order established in said block selecting step.

A method for forming a text file of characters in pixel image data comprising the steps of:

• inputting pixel image data, wherein said inputting step includes the step of converting the pixel image data into binary pixel image data in the event that the pixel image data is not binary pixel image data;
• selecting blocks of pixel image data in accordance with whether the blocks of pixel image data are text blocks or non-text blocks;
• forming a hierarchical tree of the text and non-text blocks, wherein said forming step comprises the step of designating as a descendent a connected component in the pixel image data which is within a rectangle formed around another connected component;
• grouping together connected components in the pixel image data to form text lines and text blocks;
• modifying the hierarchical tree based on the grouped-together connected components;
• segmenting text blocks into lines of pixel image data by adaptively dividing the text blocks into a least one column based on a horizontal projection of pixel density across said at least one column;
• cutting characters from the line segmented in said segmenting step, said cutting step including a first cutting layer step in which non-touching and non-overlapping characters are cut from the line segment and at least one additional cutting layer step in which touching characters are cut, said additional cutting layer step being carried out in accordance with whether or not the line characters are monospaced;
• recognizing characters cut in said cutting step and deriving character codes based on said recognition;
• and
• storing the character codes in a computer text file in accordance with an order established by the modified hierarchical tree.

A method for forming a text file of characters in pixel image data, comprising the steps of:

• inputting pixel image data, wherein said inputting step includes the step of converting the pixel image data into binary pixel image data in the event that the pixel image data is not binary image data;
• selecting blocks of pixel image data, said selecting step including the steps of outlining contours of connected components in the pixel image data, determining whether the outlined connected components include text units or non-text units based on the size of the outlined connected components, selectively connecting text units widthwisely to form text lines based on proximity of adjacent text units, and vertically connecting the formed text lines to form text blocks based on proximity of adjacent text lines and on the position of non-text units between text lines, at least one formed text block including a plurality of the formed text lines;
• forming a hierarchical tree based on the connected components outlined in said outlining step, and designating as a descendent a connected component which is within a rectangle formed around another connected component;
• modifying the hierarchical tree based on the connected text units and text lines;
• segmenting text blocks into lines of pixel data by adaptively dividing text blocks into at least one column based on a horizontal projection of pixel density across said at least one column;
• cutting characters from the lines segmented in said segmenting step, said cutting step including a first cutting layer step in which non-touching and non-overlapping characters are cut from the line segment and at least one additional cutting layer step in which touching characters are cut;
• recognizing the characters cut in said cutting step and deriving character codes based on said recognition;
• and
• storing the character codes in a computer text file in accordance with an order established by the modified hierarchical tree.

An apparatus for recognizing characters in pixel image data comprising:

• selecting means for selecting blocks of pixels from the pixel image data, said selecting means including means for outlining connected components in the pixel image data, means for forming a hierarchical tree based on the outlined connected components, means for connecting the outlined connected components to form text lines and text blocks, and means for modifying the hierarchical tree based on the connected outlined connected components, wherein said forming means comprises means for designating as a descendent a connected component which is within a rectangle formed around another connected component;
• segmenting means for segmenting text blocks into lines of pixel data by adaptively dividing the text blocks into at least one column based on a horizontal projection of pixel density across said at least one column;
• cutting means for cutting characters from the line segmented by said segmenting means, said cutting means including a first cutting means for cutting a first layer of non-touching and non-overlapping characters from the line segment and at least one additional cutting means for cutting touching characters, said additional cutting layer step being carried out in accordance with whether or not the line characters are monospaced;
• recognizing means for recognizing the characters cut by said cutting means;
• and
• storing means for storing characters recognized by said recognizing means in accordance with an order established by said block selecting means.

An apparatus for forming a text file of characters in pixel image data, comprising:

• input means for inputting pixel image data, said input means including converting means for converting the pixel image data into binary pixel image data in the event that the pixel image data is not binary pixel image data;
• selecting means for selecting blocks of pixel image data in accordance with whether the blocks of pixel image data are text blocks or non-text blocks;
• forming means for obtaining a hierarchical tree of the text and non-text blocks, wherein said forming means comprises means for designating as a descendent a connected component which is within a rectangle formed around another connected component;
• grouping means for grouping together connected components in the pixel image data to form text lines and text blocks;
• modifying means for modifying the hierarchical tree based on the grouped-together connected components;
• segmenting means for segmenting text blocks into lines of pixel image data by adaptively dividing the text blocks into at least one column based on a horizontal projection of pixel density across said at least one column;
• cutting means for cutting characters from the line segmented by said segmenting means, said cutting means including first cutting means for cutting a first layer of non-touching and non-overlapping characters from the line segment and at least one additional cutting means for cutting touching characters, said additional cutting layer step being carried out in accordance with whether or not the line characters are monospaced;
• recognizing means for recognizing characters cut by said cutting means and deriving character codes based on said recognition;
• and a computer text file in which the character codes are stored in accordance with an order established by the modified hierarchical tree.

An apparatus for forming a text file of characters in pixel image data, comprising:

• input means for inputting pixel image data, said input means including converting means for converting the pixel image data into binary pixel image data in the event that the pixel image data is not binary image data;
• selecting means for selecting blocks of pixel image data, said selecting means including outlining means for outlining contours of connected components in the pixel image data, determining means for determining whether the outlined connected components include text units or non-text units based on the size of the outlined connected components, first connecting means for selectively connecting text units widthwisely to form text lines based on proximity of adjacent text units, and second connecting means for vertically connecting formed text lines to form text blocks based on proximity of adjacent text lines and on the position of non-text units between text lines, at least one formed text block including a plurality of the formed text lines;
• forming means for forming a hierarchical tree based on the connected components outlined by said outlining means, wherein said forming means comprises means for designating as a descendent a connected component which is within a rectangle formed around another connected component;
• modifying means for modifying the hierarchical tree based on the connected text units and text lines;
• segmenting means for segmenting text blocks into lines of pixel data by adaptively dividing text blocks into at least one column based on a horizontal projection of pixel density across said at least one column;
• cutting means for cutting characters from the lines segmented by said segmenting means, said cutting means including first cutting means for cutting a first layer of non-touching and non-overlapping characters from the line segment and at least one additional cutting means for cutting an additional layer of touching characters;
• recognizing means for recognizing the characters cut by said cutting means and deriving character codes based on said recognition;
• and
• a computer text file for storing the character codes in accordance with an order established by the modified hierarchical tree.

Computer executable process steps stored on a computer readable medium, said process steps for recognizing characters in pixel image data, said process steps comprising:

• a selecting step to select blocks of pixels from the pixel image data, the selecting step including, in order, (i) an outlining step to outline contours of connected components in the pixel image data, (ii) a determining step to determine whether the outlined connected components include text units or non-text units, (iii) a forming step to form a hierarchical tree based on the outlined connected components, (iv) a first connecting step to selectively connect text units widthwisely to form text lines, (v) a second connecting step to vertically connect a plurality of text lines formed in the first connecting step to form text blocks, at least one formed text block including a plurality of formed horizontal text lines, and (vi) a modifying step to modify the hierarchical tree based on the first and second connecting steps, wherein said forming step comprises a designating step to designate as a descendent a connected component which is within a rectangle formed around another connected component;
• a segmenting step to segment text blocks into lines of pixel image data;
• a cutting step to cut characters from the line segmented in said segmenting step;
• a recognizing step to recognize characters cut in said cutting step;
• and
• a storing step to store characters recognized in said recognizing step in accordance with an order established in said block selecting step.