Contact Text Detection in Scanned Images
AssigneeSeiko Epson Corporation
InventorJing Xiao
Life of this patent
Estimating the term…
Abstract
A device and method for identifying text pixels that are erroneously classified as non-text pixels, includes accessing an image region containing a non-text component. For each non-text component within the image region component, only one determines if there are any long line structures within the bounding box defined by the non-text component. If the long line structures are greater than a predefined percentage of the span of the dimension of the bounding box parallel to the line structure, then the line structure is removed. Any remaining non-text pixels within the bounding box are reclassified as text-pixels.
Claims21
A method of reclassifying non-text pixels within a document image, comprising the following steps:
- (a) accessing said document image, said document image having foreground pixels identified, wherein foreground pixels that have been deemed to be part of human readable printable characters are classified as text pixels and foreground pixels that have been deemed not to be part of human readable printable characters are classified as non-text pixels, and connected components of text pixels are defined as text components and connected components of non-text pixels are defined as non-text components;
- (b) submitting each non-text component within said document image to a candidate selection process wherein each non-text component that is deemed to potentially include text content is designated a candidate component;
- and
- (c) submitting each designated candidate component to a reclassification process wherein foreground pixels that form a specific geometric shape are removed, and the connected components defined by any remaining foreground pixels are reclassified as text components.
The method of 1, wherein in step (c), said specific geometric shape is also of minimum size, said minimum size being defined relative to the dimensions of the bounding box of the candidate component being processed, and only foreground pixels that form the specific geometric shape and are not smaller than said minimum size are removed.
An electronic computing device for reclassifying non-text pixels within a document image, comprising:
- an input for accessing said document image, said document image having foreground pixels identified, wherein foreground pixels that have been deemed to be part of human readable printable characters are classified as text pixels and foreground pixels that have been deemed not to be part of human readable printable characters are classified as non-text pixels, and connected components of text pixels are defined as text components and connected components of non-text pixels are defined as non-text components;
- a data processor for:
- submitting each non-text component within said document image to a candidate selection process wherein each non-text component that is deemed to potentially include text content is designated a candidate component;
- and
- submitting each designated candidate component to a reclassification process wherein foreground pixels that form a specific geometric shape are removed, and the connected components defined by any remaining foreground pixels are reclassified as text components.
Dependencies are inferred from the claim text.