Abstract: Parallel processing of an image using an array of addressable registers. Image features are extracted from the image. The image features are storable as data. According to respective values of a sorting key derived from a parameter of the data, the image features are sorted into N buckets. Using an array of M addressable registers, where M is less than N, the data are summed within the buckets to perform a histogram of the image features, according to values of a histogram key derived from said a parameter of the data.

Abstract: Multiple discrete objects within a scene image captured by a single camera track are distinguished as un-labeled from a background model within a first frame of a video data input. Object position, object appearance and/or object size attributes are determined for each of the blobs, and costs determined to assign to existing blobs of existing object tracks as a function of the determined attributes. The un-labeled object blob that has a lowest cost of association with any of the existing object tracks is labeled with the label of that track having the lowest cost, said track is removed from consideration for labeling remaining un-labeled object blobs, and the process iteratively repeated until each of the track labels have been used to label one of the un-labeled blobs.

Abstract: Selection of an area of an image can be received. Selection of a subset of a plurality of predefined patterns may be received. A plurality of patterns can be generated. At least one generated pattern in the plurality of patterns may be based at least in part on one or more predefined patterns in the subset. Selection of another subset of patterns may be received. At least one pattern in the other subset of patterns may be selected from the plurality of predefined patterns and/or the generated patterns. Another plurality of patterns can be generated. At least one generated pattern in this plurality of patterns may be based at least on part on one or more patterns in the other subset. Selection of a generated pattern from the generated other plurality of patterns may be received. The selected area of the image may be populated with the selected generated pattern.

Abstract: An apparatus and method of segmenting an image using scribble segmentation is provided. An image is segmented by constraining the membership value of a subset of image elements, solving a weighted biharmonic equation subject to the constrained membership values wherein the weights are determined from similarities between image elements, and determining the final segmentation based on the membership value of each image element. An image may also be segmented by constraining a membership value of a subset of image elements, determining the unknown membership values given the constraints by solving a linear equation system using a multigrid technique, and updating a coarser level of the multigrid hierarchy to account for additional constraints using patch matrices.

Abstract: An image registration method to register a target image with a reference image is provided, including to determine a target set of features present in the target image, to determine a reference set of features present in the reference image, to generate respective signatures for each feature in the target and reference sets, to match features from the target set to corresponding features in the reference set using the signatures, to generate respective transformations to map pairs of features in the target set to corresponding pairs of features in the reference set, and determine a similarity transformation for the target image using a measure of the accuracy of the transformations.

Abstract: The disclosed methods and apparatus provide for the automatic segmentation and analysis of the overall migration rate and proliferation rate of cells that may be used with any resolution image and without the need to prepare the sample or the image before the image is analyzed. In particular embodiments, the method or apparatus of analyzing a cell image comprise performing a structure tensor analysis and/or regularization and/or a histogram-based analysis and/or a level-set analysis to classify pixels in the image into a region of interest (ROI), corresponding to cell clusters, and a non-significant region. Methods and apparatus for cell migration analysis comprise means for computing the areas of the segmented ROIs for a set of images. Methods and apparatus for cell proliferation analysis comprise means for counting the number of cells within the segmented ROIs for a set of images.

Abstract: A method of processing of photo and video images is disclosed which searches for a minimum of cost functions carried out at an N-number of image detail levels, from coarser to finer, and at each image detail level the image is divided into regions; each region is assigned a single segmentation value by a ?-number of successive iterations. The value of the cost function for the seams at the region boundaries is calculated with different types of image segmentation, and for every region a segmentation value is chosen which minimizes the sum of the cost functions of the seams and data. The technical result is the segmentation of an image with little use of the memory resources of a mobile device while maintaining both resistance to image noise and operating speed.

Abstract: A device is configured to determine a contour vector that delineates an object in an image from a remaining portion of the image, and to generate a hash value by applying a hash function to the contour vector. The device is configured to compare the hash value to a previous hash value generated by applying the hash function to a previous contour vector, where the previous contour vector is determined prior to determining the contour vector. The device is configured to determine that the hash value matches the previous hash value and, based on determining that the hash value matches the previous hash value, segment the image using the contour vector.

Abstract: A radiation therapy system includes a diagnostic image scanner (12) which acquires a multidimensional dataset of a subject that is reconstructed into at least one image representation of an object of interest. An image processing apparatus (72), of radiation therapy system, includes a segmentation unit (74) which identifies a surface contour of the object of interest, or other critical structures. A masking unit (82) determines a non-constant margin, based on the identified surface contour and appends the determined non-constant margin to the identified surface contour. The non-constant margin is based on at least one of anisotropic motion, surface morphology, positional uncertainty, proximity to other organs, and probability of dose distribution. A planning processor (70) generates a radiation therapy plan which limits the delivery of therapeutic radiation to anatomy associated with the surface contour and appended non-constant margin.

Abstract: A system and method for creating a report in a spreadsheet is disclosed. The system and method identifies a source of data items associated with the spreadsheet. The data items that are associated with the report that is generated in the spreadsheet are identified. A display including a field pane with an active tab and an all tab. The active tab displaying those data items that are determined to be associated with the report that is being generated in the spreadsheet. The all tab displaying all of the data items associated with the spreadsheet.

Abstract: Corresponding partial features are extracted from an input pattern and a registration pattern as a preregistered pattern. A plurality of combinations of the partial features are generated as composited partial features. One of registration pattern classes to which the input pattern corresponds is identified by using the plurality of composited partial features. The identification is performed based on the result of evaluation of the partial features which belong to the composited partial feature and correspond to each other in the input pattern and the registration pattern. In the evaluation, the composited partial feature is evaluated based on the measurement result of the quantity of partial features meeting a predetermined criterion.

Abstract: A stereoscopic image generation method and a stereoscopic image generation system that can generate, from an original image, a stereoscopic image that allows the viewer to perceive a natural stereoscopic effect are provided. The method includes a region setting step of setting a plurality of regions in an original image, a characteristic information acquisition step of acquiring characteristic information for each of pixels constituting the original image, a depth information generation step of generating depth information for each of the pixels for each of the plurality of regions on the basis of the characteristic information, and a stereoscopic image generation step of generating a stereoscopic image in which the positions of the pixels are changed on the basis of the pieces of depth information.

Abstract: Provided is a transition area detection device capable of detecting, with high precision, a transition area in a space without using a positioning sensor. The transition area detection device has a corresponding point search-use feature point selection unit for selecting feature points used for determining a reference image from among feature points of an input image (captured image), a geometric transformation parameter calculation-use feature point selection unit for selecting feature points used for calculating geometric transformation parameters from among feature points of the input image and feature points of the reference image, and a degree of similarity calculation-use feature point selection unit; for selecting feature points used for obtaining a degree of similarity between the captured image and the reference image from among the feature points of the input image and the feature points of the reference image.

Abstract: A method for mapping, consisting of receiving an initial set of measured values of a physiological parameter, which were measured at respective locations in a body organ and receiving a three-dimensional (3D) map of the organ including an array of spatial map elements. The method includes forming a correspondence between the respective locations at which the measured values were measured and a sub-group of the map elements, and in response to the correspondence, associating respective element values of the physiological parameter with map elements other than the sub-group. The method further includes adjusting the respective element values so that contiguous sets of the map elements form a geodesic, and displaying a map of the organ showing the adjusted element values.

Abstract: A method for identifying characters in scanned images of objects includes identifying a first set of characters in a scanned image of an object based on connected component analysis, identifying a second set of characters for the object based on an optical character recognition (OCR) process on the image of the object, and combining the first set of characters with the second set of characters to create a third set of characters.

Abstract: An image processing apparatus includes following components. A blank-area identifying unit identifies a blank area having a size exceeding a predetermined size, from a non-specifying-mark image portion of a recording area image of a recording area that includes individual images and a specifying mark specifying an area containing images among the individual images. A concealing unit conceals a divided area specified by the specifying-mark image, among divided areas obtained by dividing a non-blank area by a gap having a width equal to or larger than a threshold in a specific direction, among gaps between the individual images arranged in the specific direction in the non-blank area. The non-blank area is an area of the non-specifying-mark image portion of the recording area image other than the blank area identified by the blank-area identifying unit.

Abstract: A method of labeling pixels in an image in which pixels in the image that represent human skin of one or more people are detected and one or more regions in the image are identified, where each region in the one or more regions includes all or a portion of a human face of a person in the one or people in the image. Pixels that represent each face in the image are identified using the pixels that represent skin and the regions that include faces of the people, thereby identifying a position of each face in the image. From this, a face mask for each face and a rough body map corresponding to each face is determined using the positions of the identified faces. Further still, a torso map corresponding to each face is determined using determined face positions. Then, the extracted face masks and the torso maps are used to refine a skin map. A person or people map is determined using the skin map and the rough body map.

Abstract: A method and system for detection of native and bypass coronary ostia in a 3D volume, such as a CT volume, is disclosed. Native coronary ostia are detected by detecting a bounding box defining locations of a left native coronary ostium and a right native coronary ostium in the 3D volume using marginal space learning (MSL), and locally refining the locations of the left native coronary ostium and the right native coronary ostium using a trained native coronary ostium detector. Bypass coronary ostia are detected by segmenting an ascending aorta surface mesh in the 3D volume, generating a search region of a plurality of mesh points on the ascending aorta surface mesh based on a distribution of annotated bypass coronary ostia in a plurality of training volumes, and detecting the bypass coronary ostia by searching the plurality of mesh points in the search region.

Abstract: Hand-based biometric analysis systems and techniques are described which provide robust hand-based identification and verification. An image of a hand is obtained, which is then segmented into a palm region and separate finger regions. Acquisition of the image is performed without requiring particular orientation or placement restrictions. Segmentation is performed without the use of reference points on the images. Each segment is analyzed by calculating a set of Zernike moment descriptors for the segment. The feature parameters thus obtained are then fused and compared to stored sets of descriptors in enrollment templates to arrive at an identity decision. By using Zernike moments, and through additional manipulation, the biometric analysis is invariant to rotation, scale, or translation or an in put image. Additionally, the analysis utilizes re-use of commonly-seen terms in Zernike calculations to achieve additional efficiencies over traditional Zernike moment calculation.

Abstract: A method for separating foreground and background contents in a document image is provided. The method first computes a pixel-wise map of maximal local features (e.g., local variance, local contrast, etc.), which is binarized to generate a mask for potential foreground. In order to utilize color information effectively, the local feature map is computed using all color channels of the image. Then the background image is obtained by inpainting the mask regions from the non-mask regions of the original document image. Adaptive thresholding is applied to the difference between the original document image and the background image to obtain the binary foreground image. Post-processing of the binary foreground image can further remove undesirable elements. Finally, a more accurate background image can be obtained by inpainting the original document image using the binary foreground image as a mask.

Abstract: Techniques for identifying a salient object with respect to its context are described. A process receives an input image that includes a salient object. The process segments the input image into multiple regions and calculates a saliency value for each of the segmented regions based on scale image levels. The process constructs saliency maps based at least in part on the calculated saliency value, and combines the saliency maps to construct a total saliency map. Next, the process connects a set of line segments computed from the input image and utilizes the total saliency map to compute a closed boundary, which forms a shape prior from the closed boundary, and extracts the salient object from the total saliency map and the shape prior.

Abstract: The image signature extraction device includes an extraction unit and a generation unit. The extraction unit extracts region features from respective sub-regions in an image in accordance with a plurality of pairs of sub-regions in the image, the pairs of sub-regions including at least one pair of sub-regions in which both a combination of shapes of two sub-regions of the pair and a relative position between the two sub-regions of the pair differ from those of at least one of other pairs of sub-regions. The generation unit generates, based on the extracted region features of the respective sub-regions, an image signature to be used for identifying the image.

Abstract: An MMR system for searching across multiple indexes comprises a plurality of mobile devices, a pre-processing server or MMR gateway, and an MMR matching unit, and may include an MMR publisher. The MMR matching unit receives an image from the pre-processing server or MMR gateway and sends it to one or more of the recognition units to identify a result including a document, the page, and the location on the page. The MMR matching unit includes a distributor for distributing the image to corresponding content type specific index tables and an integrator for integrating recognition results. The result is returned to the mobile device via the pre-processing server or MMR gateway. The techniques described herein also include a number of novel methods including a method for processing content-type specific image queries and for processing queries across multiple indexes.

Abstract: An image processing apparatus has: a segmenting unit for segmenting an input image into a plurality of areas on the basis of a similarity of pixels; an acquiring unit for acquiring sizes of the segmented areas; a selecting unit for selecting one of a plurality of discriminating methods, with respect to adjacent areas among the plurality of areas, on the basis of an acquired size of each of the adjacent areas, whether or not the adjacent areas are connecting targets by using different kinds of features; an extracting unit for extracting, from the adjacent areas, the feature of the kind which is used in the selected discriminating method; a connecting unit for connecting the adjacent areas determined as the connecting targets by the selected discriminating method; and a classifying unit for classifying a kind of connected area.

Abstract: An image processing apparatus includes a detection unit configured to scan an input image and each of images at different resolutions, which are generated from the input image, by a predetermined-sized window to detect an object in the image, a storage unit configured to store a detection result of the detection unit, and a control unit configured to, if there is no free space in the storage unit to store a new detection result of the detection unit, store the new detection result instead of a detection result of an image at higher resolution than resolution of an image from which the new detection result is acquired.

Abstract: A method and an apparatus for multi-label segmentation of an image are described. In a first step user defined labels are determined for one or more pixels of the image. Then a fraction of the pixels of the image for which no user defined label is determined is pre-initialized. Finally, a cellular automaton-based segmentation of the image is performed using the user defined labels and the pre-initialized pixels.

Abstract: A method for automatically tracking multiple objects from a sequence of video images that may extract raw data about participating elements in a sporting, or other event, in a way that does not interfere with the actual participating elements in the event. The raw data may include the position and velocity of the players, the referees, and the puck, as well as the team affiliation of the players. These data may be collected in real time and may include accounting for players moving fast and unpredictably, colliding with and occluding each other, and getting in and out of the playing field. The video sequence, captured by a suitable sensor, may be processed by a suitably programmed general purpose computing device.

Abstract: An image-classification apparatus includes: a first feature extraction unit to acquire a feature value of each of block images obtained by segmenting an input image; an area-segmentation unit to assign each of the block images to any one of K areas based on the feature value; a second feature extraction unit to acquire, based on an area-segmentation result, a feature vector whose elements including, the number of adjacent spots, each including two block images adjacent to each other in the input image, for each combination of the areas whereto the two block images are assigned; or a ratio of the number of block images assigned to each of the K areas to all the number of block images adjacent to a block image assigned to each of the K areas; and a classification unit to classify to which of a plurality of categories the input image belongs.

Abstract: A system for automatically extracting or isolating structures or areas of interest (e.g., built-up structures such as buildings, houses, shelters, tents; agricultural areas; etc.) from HR/VHR overhead imagery data by way of making as little as a single pass through a hierarchical data structure of input image components (where pixels are grouped into components based on any appropriate definition or measure of dissimilarity between adjacent pixels of the input image) to identify candidate components (e.g., possible structures of interest) free of necessarily having to re-iterate the same operator configured with different threshold parameters for a plurality of values.

Abstract: Described embodiments provide an apparatus for providing a digital image by embedding recovery feature. The apparatus may include an input unit, an extraction unit, an embedding unit, and an output unit. The input unit may be configured to receive a digital image. The extraction unit may be configured to extract recovery feature from the received digital image using quad-tree decomposition. The recovery feature may be binary data that represents unique characteristics of the received digital image. The embedding unit may be configured to embed the extracted recovery feature into the received digital image. The output unit may be configured to output the digital image embedded with the recovery feature as a feature hidden image.

Abstract: A target image is deformed in such a manner that a subject in a reference image and a subject in the target image will coincide. A reference image and a target image are each divided into regions that conform to amounts of optical distortion. A region which is in the target image and which is common to the reference image and to the target image is subdivided into regions S35 to S37, S38 to S40, S42 to S44 in each of which amounts of optical distortion of both the reference image and target image are obtained. By using the amounts of optical distortion of the reference image and amounts of optical distortion of the target image 11 obtained from the subdivided regions S35 to S37, S38 to S40, S42 to S44, the target image is deformed in such a manner that a subject in the common region will coincide with the reference image.

Abstract: A people counting device includes: a person presence region extraction unit which extracts a region in which a person is present by performing person detection in relation to an input image; a person presence state estimation unit which estimates a person presence state from an extraction result; a portion detector selection unit which selects a type of portion detector to be used for people count measurement based on the estimation result, and outputs portion detector information. The people counting device further includes: a portion detection process unit which performs a portion detection process for detecting a portion of a person based on the portion detector information; and a people counting unit which obtains a number of people within the input image from a portion detection process result.

Abstract: Methods and systems for generating a shallow depth of field effect for a digitally captured image are provided. At least one region of interest (ROI) and at least one non-interest region are defined in the captured image. A difference in focus or object distance is calculated between the ROI and each non-interest region. A degree of blur is applied to each non-interest region, based on the calculated difference in focus or object distance.

Abstract: Method and apparatus for improving compression efficiency of for graphics remoting are described herein. According to one embodiment, for each data object of a stream generated within a virtual machine, the data object including graphics data representing an image to be rendered at a client over a network, it is determined whether an image associated with each data object is related to a real life content or an artificial content based on a similarity of one or more pixels selected within at least a segment of the image. A compression method is selected based on the determination of whether the image is related to a real life content or an artificial content. The image is compressed using the selected compression method and thereafter, the compressed image is streamed to the client over the network to be rendered at the client. Other methods and apparatuses are also described.

Abstract: An apparatus comprises a unit which stores a size and scene information for each of a plurality of divided areas obtained by dividing an input image, a unit which obtains a plurality of scene-based images by processing the input image based on the scene information of the plurality of divided areas, a unit which determines composite ratios of the plurality of scene-based images by determining, for each of the plurality of divided areas, a transition pattern of a composite ratio from a first composite ratio within the divided area to a second composite ratio within an area other than the divided area based on the size of the divided area, and a unit which composites the plurality of scene-based images in correspondence with the plurality of the divided areas, in accordance with the determined composite ratios.

Abstract: An image processing device includes a processor and a memory. The memory stores computer-readable instructions therein. The computer-readable instructions, when executed by the processor, causes the image processing device to perform: acquiring image data indicative of an image including an object image and a background image adjacent to the object image, the object image and the background image defining a border region in a border of the object image and the background image; acquiring at least two of a first characteristic value, a second characteristic value, and a brightness of the border region, the first characteristic value relating to a color of the object image, the second characteristic value relating to a color of the background image; and correcting a color of the border region by using the at least two of the first characteristic value, the second characteristic value, and the brightness of the border region.

Abstract: Segments included in an image I are each classified as one of object (i.e., person) segments OS1 and OS2 and foreground segments FS1 and FS2. With respect to each of the foreground segments FS1 and FS2, an importance degree is calculated based on a composition of the image I and relations between the foreground segment of the image I and a foreground segment of an image other than the image I.

Abstract: A computing system for automatically identifying individual regions in a digital image that include a human face, forming combined padded regions and automatically evaluating each of the regions including calculating a fitness score for each of the combined padded regions and ignoring at least one of the combined padded regions whose fitness score is below a known threshold.

Abstract: A system for providing an adjusted image of a scene includes an optical assembly, a capturing system coupled to the optical assembly, and a control system. The optical assembly is adjustable to alternatively be focused on a first focal area and a second focal area that is different than the first focal area. The capturing system captures a first frame of the scene when the optical assembly is focused at the first focal area, and a second frame of the scene when the optical assembly is focused at the second focal area. The first frame includes a plurality of first pixels and the second frame includes a plurality of second pixels. The control system analyzes the first frame and the second frame and utilizes graph cuts techniques to assign a depth label to at least a portion of the first frame.

Abstract: A method and system for determining a quality metric score for image processing are described including accepting a reference image, performing a pyramid transformation on the accepted reference image to produce a predetermined number of scales, applying image division to each scale to produce reference image patches, accepting a distorted image, performing a pyramid transformation on the accepted distorted image to produce the predetermined number of scales, applying image division to each scale to produce distorted image patches, performing a local distortion calculation for corresponding reference and distorted image patches, summing local distortion calculation results for image patch pairs, multiplying results of the summation operation by a positive weight for each scale, summing the results of the multiplication operation and applying a sigmoid function to results of the second summation operation to produce the quality metric score.

Abstract: Disclosed are systems and methods for automatically applying special effects based on media content characteristics. A digital image is obtained and depth information in the digital image is determined. A foreground region and a background region in the digital image are identified based on the depth information. First and second effects are selected from a grouping of effects, where the first effect is applied to at least a portion of the foreground region and the second effect is applied to at least a portion of the background region.

Abstract: An image processing apparatus that generates contour information of an image object included in input image data and compresses image data using the contour information includes a smoothing processing part that performs a smoothing process on the input image data; a contour extraction part that extracts the contour information of the image object included in the image data on which the smoothing process is performed; and a contour correction part that corrects the contour information extracted by the contour extraction part.

Abstract: A real-time stereo video signal of a captured scene with a physical foreground object and a physical background is received. In real-time, a foreground/background separation algorithm is used on the real-time stereo video signal to identify pixels from the stereo video signal that represent the physical foreground object. A video sequence may be produced by rendering a 3D virtual reality based on the identified pixels of the physical foreground object.

Abstract: CPU performs a process to obtain image data, and defines a predetermined range in the obtained image data. Further, CPU extracts outlines of images falling within the predetermined range, and selects a closed line among the extracted outlines. Then, an image processing unit and CPU converts an image surrounded by the selected closed-line into a painterly rendering image.

Abstract: A method for segmenting video data into foreground and background portions utilizes statistical modeling of the pixels. A statistical model of the background is built for each pixel, and each pixel in an incoming video frame is compared with the background statistical model for that pixel. Pixels are determined to be foreground or background based on the comparisons. The method for segmenting video data may be further incorporated into a method for implementing an intelligent video surveillance system. The method for segmenting video data may be implemented in hardware.

Abstract: The present disclosure relates to systems and methods for classifying videos based on video content. For a given video file including a plurality of frames, a subset of frames is extracted for processing. Frames that are too dark, blurry, or otherwise poor classification candidates are discarded from the subset. Generally, material classification scores that describe type of material content likely included in each frame are calculated for the remaining frames in the subset. The material classification scores are used to generate material arrangement vectors that represent the spatial arrangement of material content in each frame. The material arrangement vectors are subsequently classified to generate a scene classification score vector for each frame. The scene classification results are averaged (or otherwise processed) across all frames in the subset to associate the video file with one or more predefined scene categories related to overall types of scene content of the video file.

Abstract: An embodiment generally relates to systems and methods for estimating heart rates of individuals using non-contact imaging. A processing module can process multi-spectral video images of individuals and detect skin blobs within different images of the multi-spectral video images. The skin blobs can be converted into time series signals and processed with a band pass filter. Further, the time series signals can be processed to separate pulse signals from unnecessary signals. The heart rate of the individual can be estimated according to the resulting time series signal processing.

Abstract: Systems and methods for improving visual object recognition by analyzing query images are disclosed. In one example, a visual object recognition module may determine query images matching objects of a training corpus utilized by the module. Matched query images may be added to the training corpus as training images of a matched object to expand the recognition of the object by the module. In another example, relevant candidate image corpora from a pool of image data may be automatically selected by matching the candidate image corpora against user query images. Selected image corpora may be added to a training corpus to improve recognition coverage. In yet another example, objects unknown to a visual object recognition module may be discovered by clustering query images. Clusters of similar query images may be annotated and added into a training corpus to improve recognition coverage.

Abstract: In a method or apparatus for generating, from medical image data of a subject from a functional imaging modality, a modified intensity projection image for display. An intensity projection image data set is obtained from the image data. A segmentation for the image data is obtained. Based on information from the segmentation, a modified display property is calculated for at least one voxel of the intensity projection image data set.

Abstract: An information processing apparatus configured to generate an MR image displayed on a display apparatus includes a segmentation unit configured to segment the generated MR image into a plurality of regions, a generation unit configured to generate a pattern image to be superimposed on one of the plurality of segmented regions or each of the plurality of segmented regions, a superimposition unit configured to superimpose the generated pattern image on one of the plurality of segmented regions or each of the plurality of segmented regions, and an output unit configured to output the MR image, including the pattern image superimposed thereon, to the display apparatus.