Abstract: Classification of images or other types of high-resolution data is performed by a cluster-based data classification system. The system comprises a learner module, a classification director, and a complex classifier comprising a plurality of multi-outcome data classifiers. The classification director determines particular process rules and settings to be applied to a classification request, and the complex classifier is instantiated to process the classification request in accordance with the process rules and settings determined by the classification director. The process rules and settings are adapted under control of the learner module at least in part based on results obtained in processing the classification request and one or more additional classification requests.

Abstract: Disclosed are a road line detection method and a road line detection device. The road line detection method comprises a step of obtaining a first disparity map including one or more road regions and a corresponding V-disparity image; a step of sequentially detecting plural sloped line segments in the corresponding V-disparity image according to a big-to-small order of disparities and a big-to-small order of V-values, to serve as plural sequentially adjacent road surfaces; a step of obtaining a second disparity map of plural road line regions of interest corresponding to the plural sloped line segments; and a step of detecting one or more road lines in the second disparity map of the plural road line regions of interest.

Abstract: An image with a desired contrast is obtained while suppressing body motion artifacts caused by both random motion and periodic motion of an object. In order to do so, an imaging sequence using a non-Cartesian sampling method is executed so as to synchronize with a biological signal only at the start time and a repetition time (TR), which is an execution interval between shots within the imaging sequence, is maintained. In addition, a time difference between a delay time and a start time of each shot is calculated, and a shot with a predetermined time difference or more is executed again after the TR time.

Abstract: Embodiments of methods and/or apparatus for 3-D volume image reconstruction of a subject, executed at least in part on a computer for use with a digital radiographic apparatus can obtain a 3D volume reconstruction or projection image by generating a first-filtered set of projection images from a plurality of 2-D projection images taken over a range of scan angles and a different second-filtered set of projection images from the plurality of 2-D projection images. Then, for example, a first 3-D volume image of the subject from the first-filtered set of projection images and a second 3-D volume image of the subject from the second-filtered set of projection images can be combined using different weighting combinations in at least two corresponding portions to generate the 3-D volume image of the subject.

Abstract: A system and method for identifying a photo displays the photo with a family tree. Facial recognition, photo date estimation, people age estimation and gender determination are used to associate the displayed photo with one or more identified nodes in the family tree. A user can enter tagging information for the photo based on the identified nodes.

Abstract: Systems and methods for facilitating identification of and interaction with objects in a video frame are provided. In some embodiments, a system can include a computer-readable storage medium encoding computer executable components, and a processor that executes computer executable components encoded within the computer-readable storage medium. The components can include: a communication component that receives a video; a segmentation component that obtains a frame from the video; and a selection component that determines an object selected within the frame. The selection component can include a classifier trained using a probability map stored in the memory. The probability map can include information indicative of a likelihood that a pixel in the frame corresponds to the object, and can be generated based on crowdsourcing object differentiation.

Abstract: A control portion (140) detects pressure information fluctuations applied to a radiation sensor portion (120), acquires a radiation image and an offset image from the radiation sensor portion (120) based on the detected pressure information fluctuations, and makes offset correction of the acquired radiation image using the acquired offset image.

Abstract: Embodiments of methods and/or apparatus for 3-D volume image reconstruction of a subject, executed at least in part on a computer for use with a digital radiographic apparatus, can obtain image data for 2-D projection images over a range of scan angles. For each of the plurality of projection images, an enhanced projection image can be generated. In one embodiment, a first scatter intensity distribution through the plurality of projection images can be modulated based on a first scaling function and a SPR to generate a second scatter intensity distribution through the plurality of projection images, which can be combined with the original plurality of projection images.

Abstract: Provided is a learning device including: an acquisition section that acquires a plurality of image pairs in which the same subjects appear and a plurality of image pairs in which different subjects appear; a setting section that sets feature points on one image and the other image of each image pair; a selection section that selects a plurality of prescribed feature points, which are set at the same positions of the one image and the other image, so as to thereby select a feature extraction filter for each prescribed feature point; an extraction section that extracts the features of the prescribed feature points of each of the one image and the other image by using the plurality of feature extraction filters; a calculation section that calculates a correlation between the features; and a learning section that learns a same-subject classifier on the basis of the correlation and label information.

Abstract: An image processing apparatus to extract a print image to be printed onto a print medium from an original image, the image processing apparatus includes: a detecting unit that detects a specific area, which includes a plurality of pixels having a low degree of variation in pixel values, from the original image, based on a predetermined detection criterion; and an extracting unit that, when an extraction range having a predetermined shape including the print image is set in the original image, extracts the print image so that the specific area is disposed in a non-print area, which is not printed on the print medium, within the extraction range.

Abstract: Provided is a substrate processing apparatus that includes: a peripheral exposing unit that performs a peripheral exposing process by irradiating light to a peripheral portion of a substrate while rotating the substrate held by a substrate holding unit using a rotation driving unit; a substrate inspecting unit that performs a substrate inspecting process based on a picked up image of the substrate while moving the substrate using a movement driving unit; and a control unit. The control unit controls the predetermined substrate processing to be stopped when the peripheral exposing process is included in the predetermined substrate processing and an error occurs in the peripheral exposing unit, and controls the substrate inspecting process to be skipped when no error occurs in both of the peripheral exposing unit and a transport unit, the substrate inspecting process is included in the predetermined substrate processing and an error occurs in the substrate inspecting unit.

Abstract: A method for processing a contrast picture image of a semiconductor element. The method comprises; a color grade number reducing processing that automatically reduces number of color grades of the contrast picture image of the semiconductor element, obtained from a device for analysis, in keeping with the contrast of the contrast picture image; an interconnect contrast extraction processing that classifies pixels contained in the contrast picture image, whose number of color grades has been reduced, in accordance with a preset contrast threshold value as reference, to extract an interconnect pattern fractionated into a plurality of number of contrasts; and a shift processing that removes noise contained in a contour portion of the interconnect pattern by shifting the contour portion; whereby an interconnect pattern contained in the contrast image of the semiconductor element obtained from the device for analysis is fractionated into a plurality of preset contrasts to be extracted.

Abstract: A still image verifying apparatus includes, a storage unit that stores a plurality of pieces of summary information generated respectively from a plurality of pieces of still image data, a recording unit that records edited image data obtained by performing an edit process on still image data, and a still image decoding unit that generates decoded still image data by retrieving from the recording unit. The apparatus includes a still image determining unit that determines whether the generated decoded still image data matches the edited image data, a digest generating unit that generates summary information by retrieving from the recording unit, a digest information verifying unit that retrieves summary information corresponding to the edited image data and compares the generated summary information with the retrieved summary information and an edited still image verifying unit that verifies the authenticity of the edited image data.

Abstract: An encoding method includes selecting a run detection method in accordance with characteristics of image data, generating a run value by detecting a run that is repeated in the image data in accordance with the selected run detection method, generating run count bits by counting the number of repetitions of the detected run, and performing a run-length encoding using the detected run values and the run count bits. The run detection method includes a coefficient run detection method that detects a run based on pixel values, a pattern run detection method that detects the run based on whether the pixel blocks have a specified pattern, a difference run detection method that detects the run based on a difference value between two successive pixel values, and a residual run detection method that detects the run based on an error value that is a difference value between an actual pixel value and a predicted value.

Abstract: Provided is a method of performing an image segmentation so as to extract a building region from an input color image. The method includes extracting a plurality of straight lines that pass through a vanishing point and determining the building region in the image based on color values around the plurality of straight lines.

Abstract: Methods and Systems for training a learning based classifier and object detection in medical images is disclosed. In order to train a learning based classifier, positive training samples and negative training samples are generated based on annotated training images. Features for the positive training samples and the negative training samples are extracted. The features include an extended Haar feature set including tip features and corner features. A discriminative classifier is trained based on the extracted features.

Abstract: A method is provided in which a tag is affixed to a known individual that is to be identified within a known field of view of an image capture system. The tag is a physical tag comprising at least a known feature. Subsequent to affixing the tag to the known individual, image data is captured within the known field of view of the image capture system, which is then provided to a processor. Image analysis is performed on the captured image data to detect the at least a known feature. In dependence upon detecting the at least a known feature, an occurrence of the known individual within the captured image data is identified.

Abstract: Techniques for facilitating a shape-based search of content items are disclosed. In some situations, the techniques include receiving a request for search of content that includes an image of an item and a glyph input by the user on the image that indicates a shape of the first item, extracting features associated with the item from an area of the image on which the glyph is input; identifying a shape category of the first item based at least in part on the indicated shape of the glyph; and conducting a search for items having features that match the extracted features of the first item, the search being conducted among a plurality of items associated with the identified shape category. The retrieved search results may be returned to a user.

Abstract: Continuous wave Doppler images in a data base comprising cardiac echo studies are processed to separate Doppler frames. The frames are pre-processed to extract envelope curves and their corner shape features. Shape patterns in Doppler images from echo studies of patients with known cardiac (valvular) diseases are employed to infer the similarity in valvular disease labels for purposes of automated clinical decision support. Specifically, similarity in appearance of Doppler images from the same disease class is modeled as a constrained non-rigid translation transform of velocity envelopes embedded in these images. Shape similarity between two Doppler images is then judged by recovering the alignment transform using a variant of dynamic shape warping. Since different diseases appear as characteristic shape patterns in Doppler images, measuring the similarity in the shape pattern conveyed within the velocity region of two Doppler images can infer the similarity in their diagnosis labels.

Abstract: Systems, methods, and apparatus for sampling images using minimum mean square error subpixel-based down-sampling (MMSE-SD) are presented herein. A partition component can receive a first array of pixels, and divide the first array of pixels into two-dimensional (2-D) blocks of pixels. Further, a sampling component can diagonally down-sample subpixels of a block of the 2-D blocks, and generate a second array of pixels based on the down-sampled subpixels. The sampling component can alternately sample subpixels of adjacent pixels of the block in a diagonal direction, and generate the second array of pixels based on the subpixels. A reconstruction component can create a virtual image based on, at least in part, the second array of pixels. A MMSE-SD component can determine an optimal low resolution image based on, at least in part, respective color components of the virtual image and a high resolution image associated with the first array of pixels.