Abstract: Provided is an image processing apparatus configured to preferentially extract an image with high user satisfaction when multiple images are automatically extracted from an image group. A user specifies the number of images to be extracted from the image group. The image group is classified into a first image group with high user satisfaction and a second image group with low user satisfaction. When the number of images extracted from the first image group does not reach the specified number, an image having only a dissatisfaction factor acceptable to the user is additionally extracted.

Abstract: The application relates to particle trajectory recognition from a Centroid Population comprised of Centroids having an (x, y, t) or (x, y, f) coordinate. The method is applicable to visualization and measurement of particle flow fields of high particle. In one embodiment, the centroids are generated from particle images recorded on camera frames. The application encompasses digital computer systems and distribution mediums implementing the method disclosed and is particularly applicable to recognizing trajectories of particles in particle flows of high particle concentration. The method accomplishes trajectory recognition by forming Candidate Trajectory Trees and repeated searches at varying Search Velocities, such that initial search areas are set to a minimum size in order to recognize only the slowest, least accelerating particles which produce higher local concentrations. When a trajectory is recognized, the centroids in that trajectory are removed from consideration in future searches.

Abstract: Learning is sequentially executed with respect to weak discriminators based on learning data held in a storage device. Upon learning, an evaluation value for the weak discriminator is calculated. It is discriminated, based on a shift of the evaluation value, whether or not the learning is overlearning. If it is discriminated that the learning is overlearning, new learning data is added. Thus, the overlearning is easily detected and the learning is efficiently executed.

Abstract: In some aspects, message data is generated from a first representation of a message and a reference image. The message data is generated to hide the message such that generating a second representation of the message from the message data requires processing of a second image that is perceptually similar to the first image. In some examples, a degree to which the second representation of the message is perceptually similar to the first representation of the message is related to a degree to which the second image is perceptually similar to the first image.

Abstract: A system having a layered architecture for constructing dynamic social network from image data of actors and events. It may have a low layer for capturing raw data and identifying actors and events. The system may have a middle layer that receives actor and event information from the low layer and puts it in to a two dimensional matrix. A high layer of the system may add weighted relationship information to the matrix to form the basis for constructing a social network. The system may have a sliding window thus making the social network dynamic.

Abstract: A plurality of images (I, J) of an object (M) when viewed from different viewpoints are taken in. One of the images is set as a standard image (I), and the other image is set as a reference image (J). One-dimensional pixel data strings with a predetermined width (W) are cut out from the standard image (I) and the reference image (J) along epipolar lines (EP1, EP2) calculated from a camera parameter (CPR) and the reference point (p). Calculating a phase-only correlation function from the cut one-dimensional pixel data strings will obtain a correlation peak position (Pa1). A positional shift amount (d) from the correlation peak position (Pa1) is obtained. A search is made for a corresponding point (q) corresponding to the reference point (p) based on this position shift amount (d).

Abstract: Document data corresponding to each page included in a document is stored, and furthermore, feature data indicative of a feature of the document data and a document index indicating the document are associated with the document data. A document extracting apparatus obtains input document data, calculates feature data from the input document data, judges similarity between the input document data and the document data based on the feature data, obtains a document index associated with document data similar to the input document data, and extracts a plurality of pieces of document data associated with the document index. Thus, document data concerning the document including a page corresponding to the document data similar to the input document data is extracted for a plurality of pages.

Abstract: Systems and processes for compressing, translating and storing solid-state model data for a plurality of parts that form a solid-state model is provided. Compressing, translating and storing solid-state model data includes extracting the solid-state model data. Compressing, translating and storing solid-state model data also includes translating the sorted solid-state model data into one or more assembly components, a plurality of part components and a plurality of surface components. The process further requires storing each of the one or more assembly components, each of the plurality of part components and each of the plurality of surface components into a single file.

Abstract: An image processing apparatus includes the following elements. An image storing section stores a plurality of images related to time information. A feature extracting section extracts feature amounts of each image stored in the image storing section. An image selecting section that selects at least one of images included in each overlap portion set of several or all of the images, stored in the image storing section, on the basis of the feature amounts associated with the respective images included in the overlap portion set, the overlap portion set including overlap portions of several or all of the images overlapping with respect to shooting time. A digest-image generating section generates a digest image of the images stored in the image storing section on the basis of the selected images.

Abstract: A histogram generating function generates a histogram by counting the number of pixels included in each tone for each tone of an input video signal. A peak luminance gain calculation function calculates the gain value of a peak luminance on the basis of the histogram. A tone conversion function calculation function generates a lookup table for obtaining an output tone by correcting an input tone on the basis of the histogram and the gain value, and records the table in a tone conversion lookup table storage unit. A tone-converted image generating function which executes the correction processing of correcting the tone of an input video signal on the basis of the tone conversion lookup table.

Abstract: According to various embodiments, a method is provided for improving data and a system is provided that is configured to perform the method. The method can comprise processing a data signal by using an optical system comprising a signal processor. The method can comprise collecting data generated by the optical system, wherein the data comprises non-uniformly sampled data. The method can comprise performing an interpolation operation on the non-uniformly sampled data using the signal processor, to generate interpolated data. Further, the method can comprise adjusting the data with the interpolated data using the signal processor, to generate improved data. The improved data can be output to a user, for example, by displaying the improved data on a display unit, or by printing out the improved data. According to various embodiments, the data can comprise any desired data, for example, image data.

Abstract: Some embodiments provide a method that provides a display area for displaying an image that includes several of edges. The method provides a border drawing tool that in response to cursor movement across the image displays a search window about the cursor. The search window specifies a region to be searched to identify edges for use in defining a border for the image. In some embodiments, the size of the search window varies based on the speed of the cursor. The search window is a square box in some embodiments and a circle in other embodiments. The search window is centered at the cursor in some embodiments. In some embodiments, the display area is also for displaying the defined border over the image.

Abstract: An image processing apparatus includes: a first area extraction unit that extracts as a first area an area assuming an intensity value equal to or greater than a first threshold value from an image; a second area extraction unit that extracts as a second area an area assuming an intensity value equal to or greater than a second threshold value smaller than the first threshold value from the image; a light source area designating unit that designates as a light source area an area in the image containing a light source based upon characteristic quantities indicating characteristics of the first area and the second area; and a control unit that controls the first area extraction unit, the second area extraction unit and the light source area designating unit.

Abstract: This invention relates to method and apparatus for subpixel-based down-sampling. This invention implements an adaptive filter 140 based on edge detection, which removes visible color fringing artifacts while efficiently retaining sharpness.

Abstract: A method for providing adaptive gesture analysis may include dividing a distance range into a plurality of depth ranges, generating a plurality of intensity images for at least two image frames, each of the intensity images providing image data indicative of a presence of objects at a corresponding depth range for a respective image frame, determining motion variation between the two image frames for each corresponding depth range, and determining depth of a target based at least in part on the motion variation. An apparatus and computer program product corresponding to the method are also provided.

Abstract: An information processing apparatus recognizes an object from plural images captured by a image capture device, decides an outline of the recognized object, and calculates average luminances of the inside and the outside of the decided outline. Further, the information processing apparatus determines that a difference between the average luminances of the inside and the outside of the outline is equal to and more than a predetermined value, generates an adjustment image which make gradation increase so that a luminance of any one of the inside and the outside of the outline increases when the difference is equal to and more than the predetermined value, the any one of the inside and the outside of the outline having a lower average luminance than another one. The adjustment image is projected onto the object by a projection device, and then captured by the image capture device.

Abstract: A moving-picture coding device that performs inter-field motion compensation prediction, includes a reference limit area determining unit that determines to select, for each reference picture, with respect to reference pictures used for coding, a luminance or chrominance reference limit area putting a limit to a reference picture used for generating a prediction picture of a luminance or chrominance of a coding target block belonging to the coding target picture limit area of the case upon each reference picture being previously coded; a selecting inhibit vector determining unit that determines, in a case of coding the coding block belonging to the coding target picture limit area, a selecting inhibit vector that inhibits selecting a luminance and the chrominance are comprised only of pixels of the luminance reference limit area and the chrominance reference limit area.

Abstract: A method of determining a regular grid pattern from a surface coded pattern that comprises the regular grid pattern interleaved with a further data carrying pattern wherein the surface coded pattern is subject to perspective distortion, the method comprising: extracting a set of straight line hypotheses from the coded surface pattern; clustering the straight line hypotheses by orientation; for each cluster, extracting a set of line pencil hypotheses; generating a set of regular grid hypotheses from pairs of the line pencil hypotheses; and determining the best regular grid hypothesis.

Abstract: Aspects of the present invention are related to systems and methods for connected-component labeling.

Abstract: A method and system for brain tumor segmentation in multi-spectral 3D MRI images is disclosed. A trained probabilistic boosting tree (PBT) classifier is used to determine, for each voxel in a multi-spectral 3D MR image sequence, a probability that the voxel is part of a brain tumor. The brain tumor is then segmented in the multi-spectral 3D MRI image sequence using graph cuts segmentation based on the probabilities determined using the trained PBT classifier and intensities of the voxels in the multi-spectral 3D MR image sequence.