Sharp Laboratories of America, Inc

Lawrence Shao-hsien Chen

Richard John Campbell

Toyohisa Matsuda

Ahmet Mufit Ferman

John Edward Dolan

Ching-Wei Chang

Ronnie Neil Patton

Jon Matthew Speigle

Aspects of the present invention are related to systems and methods for automatically determining the content boundaries in a digital image. A gradient field may be generated using an edge detector, and the gradient field may be thresholded, by magnitude, to retain strong edges. The resulting localized edge positions may be projected onto a first direction and a second direction to form two projection histograms. The projection histograms may be analyzed to determine the boundaries of the image content. Corners of a cropping rectangle may be computed, and the digital image may be cropped according to the content boundaries.

Methods and systems for content-boundary detection

Aspects of the present invention relate to methods and systems for fitting a mesh to an image of a document page and correcting a distortion in the image of the document page using the mesh. A rectangular-form mesh may be overlaid on a displayed document image. Corner points in the rectangular-form mesh may be moved to coincide with corner points of a document page in the displayed document image, and a corner-aligned mesh may be overlaid on the displayed document image. A cubic Bezier curve may be fit to the top edge of the document page, a cubic Bezier curve may be fit to the bottom edge of the document page and an edge-aligned mesh may be overlaid on the displayed image. An inverse projective transformation may be applied to the patches of the mesh to correct the distortion in the image of the document page.

Methods and systems for correcting a document image

Aspects of the present invention are related to systems and methods for providing real-time feed-back cues in an image acquisition process. An overlaid preview-frame image may be rendered comprising cues associated with acquisition factors. A cue may indicate whether, or not, an acquisition-factor value falls within a range of acceptable values. Image features may be extracted, from a preview-frame image, and interpreted to determine acquisition-factor values and their acceptability.

Methods and systems for real-time image-capture feedback

Aspects of the present invention are related to systems and methods for correcting artifacts in a camera-captured image of a document or image of an object exhibiting document-like content. A mobile device may capture an image and send the image to a cloud computing system for processing. According to a first aspect of the present invention, the mobile device may provide real-time feedback cues to assist in the capture of an image. The mobile device may detect a region-of-interest in the captured image, and a user may refine or confirm the detected region-of-interest. The captured image, information identifying the region-of-interest and a metadata tag, referred to as a region-of-interest modification tag, indicating whether, or not, the region-of-interest was refined by a user may be sent to the cloud. The cloud may process the image giving priority to the region-of-interest received from the handset when the region-of-interest modification tag indicates that the region-of-interest was refined by a user over a cloud determined region-of-interest. The cloud may transmit, to the handset, the processing results.

Methods and systems for mobile document acquisition and enhancement

Aspects of the present invention are related to systems and methods for automatic content-boundary detection in a digital image. According to one aspect of the present invention a received image may be preconditioned to form a normalized image from which image-content edges may be detected. Projection histograms in the direction of a known skew angle associated with the image and the normal to the skew angle may be formed to determine image-content boundaries.

Methods and systems for automatic content-boundary detection

Aspects of the present invention are related to systems and methods for identifying a document region in a non-scanned image. A first candidate document region-of-interest mask may be determined from a luminance-only image by foreground elimination, non-uniformity reduction and thresholding. A second candidate document region-of-interest mask may be formed by subtracting from the first candidate document region-of-interest mask an edge mask resulting from detecting edges in the luminance-only input image. The second candidate document region-of-interest mask may be refined, to form a final document region-of-interest mask, based on text features calculated from the luminance-only input image. The document region-of-interest may be extracted using the final document region-of-interest mask.

Methods and systems for determining a document region-of-interest in an image

Embodiments of the present invention comprise systems, methods and devices for detection of pictorial regions in an image using a masking condition, an entropy measure, and region growing.

Methods and systems for detecting pictorial regions in digital images

Aspects of the present invention are related to systems and methods for correcting illumination and vignette in a camera-captured document image. A background type may be detected from down-sampled luminance image associated with the camera-captured image, and model parameters may be estimated for a morphologically filtered version of the down-sampled luminance image, wherein the morphological filter may be based on the detected background type. The model parameters may be verified and if deemed acceptable, a rectified image may be formed.

Methods and systems for document-image correction

Embodiments of the present invention comprise methods and systems for identifying uniformly colored regions in a digital image.

Methods and systems for identifying regions of substantially uniform color in a digital image

Aspects of the present invention relate to methods and systems for segmenting a digital image into regions. A frequency-of-occurrence of image values may be determined excluding a portion of pixels in a digital image. An image value associated with a peak in the frequency-of-occurrence of image values may be identified and associated with an image label. Pixel locations may be labeled based on the associated labels and image values. Additionally, unreliable pixels may be determined and labeled based on the associated labels and image values, and an unreliable pixel may be assigned a label after multiple scan passes of a classification map.

Methods and systems for segmenting a digital image into regions

A software-controlled method for minimizing color toner use during copying of a color page by a color copier including (a) examining such a page to identify different pre-selected categories of page-contained image elements, (b) associating, with identified different-categories of such elements, specific, respective, pre-determined color-toner usage rules, and (c) following the associating step, outputting the elements to the copier along with instructions that the copier use, in the copying of such elements, the correct, respectively associated color-toner usage rule(s). From a structural point of view, the invention includes examining structure, associating structure, and output structure designed to perform the basic core methodology steps of the invention.

Intelligent toner saving for color copying

Embodiments of the present invention comprise systems and methods for refining text-detection results for a digital image.

Methods and systems for refining text detection in a digital image

Aspects of the present invention relate to systems, methods and devices for detection of text in an image using an initial text classification result and a verification process.

Methods and systems for identifying text in digital images

A method and apparatus for determining an orientation of a document including Korean text are presented. A binarized pixel image is created from the document image. Contiguous pixels are grouped and labeled using a bounding box. A spanning stroke may be detected from a group of the contiguous pixels. The orientation of the document is determined by comparing counts associated with spanning strokes in the left, right, top, and bottom halves of the bounding boxes.

Method and apparatus for determining an orientation of a document including Korean characters

Aspects of the present invention relate to systems, methods and devices for detection of text in an image using an initial text classification result and a verification process. In particular, a support region of a candidate text pixel in a text-candidate map may be expanded to produce a revised text-candidate map. Pictorial regions in the image may be discriminated based on an entropy measure using masking and the revised text-candidate map, and the revised text-candidate map may be refined based on the pictorial regions.

Growing community of inventors

Vancouver, WA, United States of America

Lawrence Shao-hsien Chen