Abstract: An image processing apparatus analyzes an image of a document to thereby extract a heading region from the image. The image processing apparatus detects candidates for the heading region from the image, and defines a predetermined range in the image as a range to be processed. The apparatus further groups the candidates in the range to be processed, based on a feature quantity corresponding to a feature in terms of style of a character string. The apparatus selects a representative group from the resultant groups, and divides the range to be processed, at the position of a candidate belonging to the representative group. The apparatus newly defines each of the portions generated by the division as a range to be processed.

Abstract: A color that readily leaves an impression is adopted as the representative color of a color image. Pixels that form a color image are distributed in L*a*b* color space and are projected onto the a*b* plane. The pixels are grouped and the color at the centroid position of each group is adopted as a representative candidate color. A first score is calculated based upon the distance from the origin, which is indicative of gray, to the representative candidate color, and a second score is calculated from the number of pixels contained in the group. A final score is calculated from the first and second scores and representative candidate colors that provide the three highest final scores are decided upon as representative colors.

Abstract: A method, system and computer software product for improving rate-distortion performance while remaining faithful to JPEG/MPEG syntax, involving joint optimization of Huffman tables, quantization step sizes and quantized coefficients of a JPEG/MPEG encoder. This involves finding the optimal coefficient indices in the form of (run, size) pairs. By employing an interative process including this search for optimal coefficient indices, joint improvement of run-length coding, Huffman coding and quantization table selection may be achieved. Additionally, the compression of quantized DC coefficients may also be improved using a trellis-structure.

Abstract: A color image comprises color values in each of one or more color channels for each of a plurality of points, or pixels, within the image. The image is represented by rank ordering the values in each color channel. The image representation generated in this way is usable for automated-vision or computer-vision tasks, for example.

Abstract: The present invention is a method and system for measuring human emotional response to visual stimulus, based on the person's facial expressions. Given a detected and tracked human face, it is accurately localized so that the facial features are correctly identified and localized. Face and facial features are localized using the geometrically specialized learning machines. Then the emotion-sensitive features, such as the shapes of the facial features or facial wrinkles, are extracted. The facial muscle actions are estimated using a learning machine trained on the emotion-sensitive features. The instantaneous facial muscle actions are projected to a point in affect space, using the relation between the facial muscle actions and the affective state (arousal, valence, and stance). The series of estimated emotional changes renders a trajectory in affect space, which is further analyzed in relation to the temporal changes in visual stimulus, to determine the response.

Abstract: A method, medium, and system compressing an image, and a method, medium, and system recovering an image. Values of colors of a pixel from among pixels making up an image are predicted from values of colors of a reference pixel corresponding to the pixel, and the predicted values of the colors of the pixel are corrected based on similarities in variations in color values in the image.

Abstract: A method and device is provided for recognizing characters in a handwritten input representing an input character string. A character sub-string preceding an unrecognized character in the input character string is determined. Handwriting recognition is used to provide one or more candidate characters for the unrecognized character. One of the one or more candidate characters is then selected. The candidate character selected, is the one which is most likely to be a correct recognition of the unrecognized character based on the determined character sub-string.

Abstract: A color-reproduction system is provided. The system includes an image-content-analysis unit that analyzes the color distribution of pixels that constitute an inputted image, a parameter-prediction unit that predicts scaling parameters based on the attribute information on the pixels, the color distribution, and the image size information, and a luminance-chroma-determination unit that determines the final luminance and chroma of the pixels using the predicted parameters.

Abstract: The geometry of an object is inferred from values of the signed distance sampled on a uniform grid to efficiently model objects based on data derived from imaging technology that is now ubiquitous in medical diagnostics. Techniques for automated segmentation convert imaging intensity to a signed distance function (SDF), and a voxel structure imposes a uniform sampling grid. Essential properties of the SDF are used to construct upper and lower bounds on the allowed variation in signed distance in 1, 2, and 3 (or more) dimensions. The bounds are combined to produce interval-valued extensions of the SDF, including a tight global extension and more computationally efficient local bounds that provide useful criteria for root exclusion/isolation, enabling modeling of the objects and other applications.

Abstract: A method for matching colors including comparing the appearance of a first white color associated with a first color imaging system and a second white color associated with a second color imaging system, wherein the tristimulus values of the first and second white color are similar; determining a fixed correction to the tristimulus values of the second white color to achieve a visual match to the first white color; measuring a first set of spectral values for a first color associated with the first color imaging system; determining a first set of tristimulus values from the first set of spectral values; measuring a second set of spectral values for a second color associated with the second color imaging system; determining a second set of tristimulus values from the second set of spectral values; applying a correction to the tristimulus values of the second color; determining a difference between the tristimulus value of the first color and the corrected tristimulus value of the second color; and adjusting th

Abstract: An image processing apparatus includes a separation unit which determines the attribute of data contained in input document image data and separates the document image data into areas by attributes, an extraction unit which extracts, from the separated areas, an area of a graphics image as a target of vectorization processing, a determination unit which determines whether the attribute of the area of the graphics image is a clipart area or a line drawing area including a line drawing, and a vector conversion unit which performs vectorization processing corresponding to the attribute of the graphics image based on the determination result of the determination unit.

Abstract: An image encoding device includes a single image pickup unit generating image data by shooting a subject, an image compression unit generating digital image data by encoding the image data, a distance measuring unit measuring a distance to the subject during the shooting, and a multiplexing unit generating multiplexed data based on the digital image data and the measured distance by multiplexing distance data indicating the distance to a subject of each of given pixel blocks of the digital image data onto the digital image data.

Abstract: A method for matching colors includes providing a first source of a first color and a second source of a second color; measuring a set of spectral values for first color stimuli associated with the first color; calculating a first set of tristimulus values; defining a first set of human observer color matching functions; calculating a first set of corrected tristimulus values from the first set of spectral values using the first set of human observer color matching functions; measuring a second set of spectral values for second color stimuli associated with the second color; calculating a second set of tristimulus values; defining a second set of human observer color matching functions; calculating a second set of corrected tristimulus values from the second set of spectral values; determining a difference between the first and second set of corrected tristimulus values; and adjusting one of the sources to reduce the difference.

Abstract: An image processing apparatus includes: a reference-based coding unit that encodes image information for an image partition having a predefined size by referring to image information for another image partition; an independently coding unit that encodes the image information for the image partition independently of any other image partition; and a bounds defining unit that defines bounds of reference to be made by the reference-based coding unit.

Abstract: A method of text extraction in color compound documents is described. The method connects similarly colored pixels of an image of a color compound document into connected components (CCs); classifies each CC as either text or non-text; refines the text CC classification for each text CC using global color context statistics; groups text CCs into text blocks; recovers misclassified non-text CCs into a nearby text block; and removes extraneous CCs from each text block using local color context statistics to thereby provide the extracted text in the text blocks. Also described is a method of locating graphics objects in a color compound document image.

Abstract: A region separation unit separates an inputted color document image into a plurality of types of regions such as a character region, a clip art region and a photo image region, and a clip art region extraction unit identifies the clip art region from among the separated regions. A clip art region dividing unit divides the clip art region based on the color features of the clip art region, and a clip art background identify unit identifies the background portion of the clip art region from among the divided regions. A filling unit for filling portions other than the background of a clip art fills a portion of the clip art other than the background with the background color, and a JPEG compression unit compresses the result obtained from the process for filling a clip art portion.

Abstract: An automotive display device, which allows the user to easily recognize the sameness of an information image even if the shape of the image is changed, is realized. An automotive display device (1) includes a display section (13) for displaying a vehicle information image including information regarding a vehicle. The vehicle information image is a 3-D image. The automotive display device (1) further includes an image processing section (113) for generating a vehicle information image seen from a fixed viewpoint, an image display processing section (114) for displaying the vehicle information image on the display section (13), and an ETC information detection section (21) that detects a driving condition of the vehicle. The image processing section (113) rotates the vehicle information image around a predetermined axis and for a predetermined angle, in accordance with ETC information obtained by the ETC information detecting section (21).

Abstract: A method for encoding an image is disclosed. The method generally includes the steps, of (A) generating a quantization matrix as a function of at least four parameters, (B) optimizing the parameters to maximize a quality metric for encoding the image and (C) encoding the image with the quantization matrix as optimized.

Abstract: Techniques, apparatus and systems for image decoding are described. A method performed by an image decoding apparatus includes generating motion compensation information and a quantized result of an error image that represents a difference between a current image to be decoded and a prediction image obtained by performing inter prediction on the current image from a bitstream of the current image. A reference image of the current image is received from an external memory unit based on the generated motion compensation information. The method includes restoring the current image based on the received reference image of the current image and the generated quantized result of the error image representing the difference between the current image and the prediction image. Receiving the reference image of the current image is performed while restoring a previous image, and restoring the current image is performed while receiving a reference image of a subsequent image.

Abstract: A DCT-based technique with rhombus scanning for image compression. A flipped-kernel discrete cosine transform is applied to an eight by eight pixel sub-block of the sixteen by sixteen pixel block. A visually insignificant information is removed from the eight by eight pixel sub-block. A quantization method is used to remove the visually insignificant information. A quantized discrete cosine transform coefficient is scanned of the sixteen by sixteen pixel block. The quantized discrete cosine transform coefficient is scanned according to a rhomboid pattern. A portion of a digital image may be divided into a sixteen by sixteen pixel block.