Abstract: An image recognition apparatus includes a reception part that receives an image that has been read; a determination part that determines a registered object to correspond to an object included in the received image that has been read from among previously registered plural objects; a reflecting part that reflects colors of the image that has been read in previously stored plural similar objects each similar to the registered object determined by the determination part; and a printing control part that causes a printing apparatus to print the plural similar objects in which the colors have been reflected by the reflecting part.

Abstract: An automated method for texture segmentation (11) of geophysical data volumes, where texture is defined by double-window statistics of data values, the statistics being generated by a smaller pattern window moving around within a larger sampling window (12). A measure of “distance” between two locations is selected based on similarity between the double-window statistics from sampling windows centered at the two locations (13). Clustering of locations is then based on distance proximity (14).

Abstract: A computer-assisted method of classifying cytological samples, includes using a processor to analyze images of cytological samples and identify cytological objects of interest within the sample images, wherein the processor (i) displays images of identified cytological objects of interest from the sample images to a reviewer, (ii) accesses a database of images of previously classified cytological objects, and (iii) displays to the reviewer, interspersed with the displayed images of the identified objects of interest from the sample images, one or more images obtained from the database of images of previously-classified objects.

Abstract: The disclosure is directed to techniques for region-of-interest (ROI) processing for video telephony (VT) applications. According to the disclosed techniques, a recipient device defines ROI information for video information transmitted by a sender device, i.e., far-end video information. The recipient device transmits the ROI information to the sender device. Using the ROI information transmitted by the recipient device, the sender device applies preferential encoding to an ROI within a video scene. ROI extraction may be applied to process a user description of a region of interest (ROI) to generate information specifying the ROI based on the description. The user description may be textual, graphical, or speech-based. An extraction module applies appropriate processing to generated the ROI information from the user description. The extraction module may locally reside with a video communication device, or reside in a distinct intermediate server configured for ROI extraction.

Abstract: An image processing device, method and program in which a feature point derivation unit derives a plurality of characteristic points in an input moving image. A tracking subject feature point setting unit sets a feature point within a tracking subject, from the characteristic points. A background feature point setting unit sets a group of background feature points from the characteristic points. The background feature points are not located within the tracking subject. A motion detection unit detects movement over time of the background feature points. A clip area setting unit sets a size and a position of a clip area of an image to be employed which includes the feature point within the tracking subject, on the basis of the movement of the feature point within the tracking subject and the movement of the background feature points, when the motion detection unit detects movement of the background feature points.

Abstract: A target recognition system operatively connected to a stereo imaging device to capture a stereo image of an area ahead of the target recognition system, includes a parallax calculator to calculate parallax of the stereo image including two captured images; a target candidate detector to detect a candidate set of recognition target areas based on a luminance image of one of the captured images; and a target recognition processor to limit the candidate set of recognition target areas detected by the target candidate detector within a range of threshold values of characteristics in the candidate set of recognition target areas set in advance based on the parallax calculated by the parallax calculator to extract and output the one or more recognition targets.

Abstract: A system and method includes data representing a sequence of X-ray images of a portion of patient anatomy acquired over a time interval and signal data representing electrical activity of the heart of the patient over the time interval, determination of a score value for each image of said sequence of X-ray images, selection of a set of images from said sequence of X-ray images based on the determined score values, the set of images excluding one or more images of said sequence of X-ray images, and generation of an averaged image from said set of images.

Abstract: Provided is a carried item region extraction device for accurately extracting a carried item region from an image. This carried item region extraction device has: a string region processing unit for extracting a string region including a string of a carried item from image information; and a carried item region processing unit for extracting a carried item region including a carried item from the image information on the basis of the string region.

Abstract: A computer implemented system for identifying license plates and faces in street-level images is disclosed. The system includes an object detector configured to determine a set of candidate objects in the image, a feature vector module configured to generate a set of feature vectors using the object detector to generate a feature vector for each candidate object in the set of candidate objects, a composite feature vector module to generate a set of composite feature vectors by combining each generated feature vector with a corresponding road or street description of the object in question, and an identifier module configured to identify objects of a particular type using a classifier that takes a set of composite feature vectors as input and returns a list of candidate objects that are classified as being of the particular type as output.

Abstract: An encoding method for encoding of images and corresponding decoding method is disclosed, as well as an encoder and a decoder. The encoding method comprises creating, in an encoder, a residual block from an original block and a prediction block which is an approximation of the original block, wherein the created residual block comprises residual block elements. The en-coding method further comprises re-ordering, in the encoder, the residual block elements within the residual block; and applying, in the encoder, a transform to the re-ordered residual block. By re-ordering the residual block elements prior to applying a transform, an improved efficiency of the encoding may be obtained.

Abstract: A method for enhancing a three-dimensional (3D) image comprising at least two depth layers wherein each depth layer comprising image objects. The method comprising the steps of determining a near field and a far field comprising at least one depth layer each, identifying the image objects in the near field and the far field respectively, applying a first correction curve to the image objects identified in the near field and applying a second correction curve to the image objects identified in the far field.

Abstract: A target recognition system and a target recognition method to recognize one or more recognition targets, operatively connected to an imaging device to capture an image of an area ahead of the target recognition system, each of which includes a recognition area detector to detect multiple recognition areas from the captured image; a recognition weighting unit to set recognition weight indicating existence probability of images of the recognition targets to the respective recognition areas detected by the recognition area detector; and a target recognition processor to recognize the one or more recognition targets in a specified recognition area based on the recognition weight set in the respective recognition area.

Abstract: A behavior analysis device has an object extraction portion that processes a frame image of an imaging area being imaged by an imaging device and extracts an object being imaged, a position detection portion that detects a position in the imaging area, for each object extracted by the object extraction portion, a posture estimation portion that estimates an posture of the object, for each object extracted by the object extraction portion, and a behavior determination portion that determines a behavior of the object, for each object extracted by the object extraction portion, based on the position in the imaging area that is detected by the position detection portion and the posture estimated by the posture estimation portion.

Abstract: More than two acquired energy or spectral M bins are used for photon counting detectors in a CT system. In the pre-reconstruction data decomposition, the measured photo counts in the M acquired spectral bines are combined into a predetermined fewer number of processed or weighted spectral bines N, which is at least two in number and represents a number of selected basic materials. Since the N basis materials are selected in the imaged subject where N<M, the noise is balanced in the pre-reconstruction data decomposition. In some more detail, after the photon counts n(E) is acquired in each of the M spectral bins as indicated by E=1 to M at each of the detector units for every view, the M acquired spectral bins is combined into the N processed spectral bins so that the photon counts in the N processed spectral bins are substantially balanced to have a least number of differences among the N processed spectral bins.

Abstract: A learning apparatus in the present invention includes a weak discriminator generation unit that generates a weak discriminator which calculates a discrimination score of an instance of a target based on a feature and a bag label, a weak discrimination unit which calculates the discrimination score based on the generated weak discriminator, an instance probability calculation unit that calculates an instance probability of the target instance based on the calculated the discrimination score, a bag probability calculation unit that calculates a probability that no smaller than two positive instances are included in the bag based on the calculated instance probability and a likelihood calculation unit which calculates likelihood representing plausibility of the bag probability based on the bag label.

Abstract: An object recognition unit recognizes, from real-space video data, a body included in the video data. A function setting unit retains function information in which is prescribed a function configured from a pair of operation and processing that can be set for each type of body. In addition, the function setting unit sets, to each body recognized by the object recognition unit, a function that can be set, based on the type of each body. A selection determination unit determines a selected body selected by a user as the body to be operated among the respective bodies recognized by the object recognition unit. An operation determination unit determines the operation that the user has performed on the selected body. A processing determination unit 1107 determines the processing for the operation that has been determined by the operation determination unit among the operations configuring the function set by the function setting unit.

Abstract: Disclosed are methods and apparatus for inspecting an extreme ultraviolet (EUV) reticle is disclosed. An optical inspection tool is used to obtain a phase defect map for the EUV reticle before a pattern is formed on the EUV reticle, and the phase defect map identifies a position of each phase defect on the EUV reticle. After the pattern is formed on the EUV reticle, a charged particle tool is used to obtain an image of each reticle portion that is proximate to each position of each phase defect as identified in the phase defect map. The phase defect map and one or images of each reticle portion that is proximate to each position of each phase defect are displayed or stored so as to facilitate analysis of whether to repair or discard the EUV reticle.

Abstract: A defect inspection method comprising: picking up an image of a subject under inspection to thereby acquire an inspection image; extracting multiple templates corresponding to multiple regions, respectively from design data of the subject under inspection; finding a first misregistration amount between the inspection image and the design data using a first template as any one template selected from among the plural templates; finding a second misregistration amount between the inspection image and the design data using a second template other than the first template, the second template being selected from among the plural templates, and the first misregistration-amount; and converting the design data, misregistration thereof being corrected using the first misregistration-amount, and the second misregistration-amount, into a design data image, and comparing the design data image with the inspection image to thereby detect a defect of the subject under inspection.

Abstract: According to one embodiment, it is provided that an image compressor includes an image data controller, first to third encoders and an encoded data generator. The image data controller extracts first to third pixels. The first encoder encodes a target first pixel and generates a first encoded pixel. The second encoder encodes a target second pixel and generates a second encoded pixel. The third encoder encodes a target third pixel and generates a third encoded pixel. The encoded data generator combines the first to third encoded pixels and generates encoded data.

Abstract: In one aspect, a computer implemented method of motion capture, the method includes tracking the motion of a dynamic object bearing a pattern configured such that a first portion of the patterns is tracked at a first resolution and a second portion of the pattern is tracked at a second resolution. The method further includes causing data representing the motion to be stored to a computer readable medium.