Abstract: An optical processing system comprises at least one spatial light modulator, SLM, configured to simultaneously display a first input data pattern (a) and at least one data focusing pattern which is a Fourier domain representation (B) of a second input data pattern (b), the optical processing system further comprising a detector for detecting light that has been successively optically processed by said input data patterns and focusing data patterns, thereby producing an optical convolution of the first and second input data patterns, the optical convolution for use in a neural network.

Abstract: A method for extracting signatures and assigning signatory name attributions to the extracted signature images from unstructured digital documents is provided. The method includes: receiving a document that includes a signature; detecting a first region within the document in which the signature is present; applying a region-based convolutional neural network to the detected first region in order to demarcate a boundary that surrounds the detected first region; detecting a second region within the document in which a name that relates to the signature is present, by scanning the document to obtain a set of text words and then applying a named entity recognition (NER) machine learning algorithm to determine which text words are names; and assigning a signatory name attribution to the name that is present in the second region, based on a calculated distance between the name and the demarcated boundary.

Abstract: Exemplary methodology, procedure, system, method and computer-accessible medium can be provided for authenticating a non-digital medium of a physical object, by receiving at least one image of video of at least one marked or unmarked region, and comparing the first microscopic image or video of at least one marked or unmarked region with at least one second microscopic image or video relating to the non-digital medium to determine if a similarity between the first and second microscopic images or videos matches or exceeds a predetermined amount.

Abstract: An information processing apparatus includes a detector, a receiver, and an image processor. The detector is configured to detect a punch hole candidate partial image having a punch hole accuracy indicating whether the punch hole candidate partial image is a punch hole image is less than a first threshold value and equal to or greater than a second threshold value smaller than the first threshold value, from plural partial images constituting a read image of a document. The receiver is configured to display the punch hole candidate partial image, and receive an instruction from a user as to whether to perform image processing on the punch hole candidate partial image. The image processor is configured to perform the image processing on the punch hole candidate partial image when the receiver receives an instruction to perform the image processing on the punch hole candidate partial image.

Abstract: A specification of an expected border of a bounded area is received. One or more images of at least a portion of the bounded area that is at least a threshold distance away from the expected border are received to generate a model of the bounded area. A current location position of the aerial vehicle is used to determine that the aerial vehicle is within the threshold distance away from the expected border. In response, an updated expected border is determined using the generated model of the bounded area and a border image of at least a portion of the expected border captured by an image sensor.

Abstract: An image processing apparatus includes a punch hole trace detection unit that detects a punch hole trace from an input image, a line segment detection unit that detects a line segment which is positioned in a vicinity of the detected punch hole trace, and a punch hole trace processing unit that performs processing on the punch hole trace on the basis of a result of the detection by the line segment detection unit.

Abstract: Techniques are provided for detection and location of active display regions in videos with static borders. A methodology implementing the techniques according to an embodiment includes extracting features from rows and columns of pixels of a video frame. The features are based on horizontal gradient runs (HGRs) and vertical gradient runs (VGRs). The method also includes detecting one or more static regions of the frame, based on a comparison of differences between the features of the current video frame and features extracted from a previous video frame. The method further includes detecting one or more boundaries of the static regions based on a location of a maximum value of one of the features within the static region, if the maximum value is greater than a boundary detection threshold value. Determination of the active region in the current video frame is based on exclusion of the detected static regions.

Abstract: Systems and methods are described for augmenting video data based on automated identification of one or more objects depicted in the video data. One or more classification models may identify an object of interest in video data. An aggregated duration count may be maintained that reflects a length of time that the object of interest has been depicted in the video data. This duration or additional metric data derived in part from the duration may be displayed in association with display of the video data and continuously updated during playback of the video data.

Abstract: A method and device for detecting connected pixels in a two-dimensional (2D) image having a plurality of pixels each having X-Y coordinates may compare an upper-row pixel included in a Yth row with a lower-row pixel included in a Y+1th row depending on a comparison condition, wherein Y is a positive integer, and wherein when a difference in X coordinate between the upper-row pixel and the lower-row pixel is one or less, the upper-row pixel and the lower-row pixel are determined to be positioned adjacent each other, and when the difference in X coordinate is two or more, the upper-row pixel and the lower-row pixel are determined to be spaced apart from each other and update an index map that defines a relationship between index numbers of adjacent pixels depending on the determination to set pixels with the same or associated index number in the same pixel group.

Abstract: A first area, which includes a group of pixels with a first color, is specified. The first color corresponds to a color of an outline of a hole image in an image represented by image data. A second area corresponding to the hole image is determined. The second area is the first area larger than a first size when the first area is separated from four ends of the image. The second area is the first area larger than a second size when the first area is in contact with at least one of the four ends of the image. The second size is smaller than the first size.

Abstract: A display device and a method of driving the same are disclosed. In one aspect, the display device includes a display panel including a plurality of pixels including a first group of pixels and a second group of pixels. The first group of pixels forms a first region and the second group of pixels forms a second region surrounding the first region. A controller is configured to receive input image data, process the input image data corresponding to the first pixels based on a preset first image processing algorithm so as to generate first modified image data, and process the input image data corresponding to the second pixels based on a preset second image processing algorithm so as to generate second modified image data.

Abstract: A clustering processing unit sets representative data to be used in first clustering, for each of temporally continuous images, and repeats clustering of pixels based on the representative data. A representative data updating unit updates the representative data whenever clustering is repeated. As the representative data to be used in first clustering of each image, an initial value or representative data used in last clustering of an immediately preceding image is set based on predetermined input information. An example of the predetermined input information is operation information for an image capturing unit for capturing an image. If a dead cluster predictively forms due to an image change, the initial value is set as the representative data. This makes it possible to restrict accumulation of dead clusters, and suppress the decrease in segmentation accuracy.

Abstract: Character recognition is described. In one embodiment, it may use matched sequences rather than character shape to determine a computer-legible result.

Abstract: Identifying whether an image includes a search image can be accomplished using region analysis and/or parts model analysis. Region analysis can include dividing a received image into regions and analyzing each region by: enlarging the region, applying a gradient histogram transformation to the enlarged region, and comparing the transformed region, at various angles, to a set of search patterns. Parts model analysis can include identifying key features of an image such as prominent lines or areas, objects or object edges, consecutive colors; generating one or more parts models from the key features by representing the key features as geometric objects; and comparing each parts model, at various angles, to a set of search patterns. The comparison employed in region analysis or parts model analysis can employ classifiers trained on the search images.

Abstract: This disclosure describes systems, and methods related to determining carrier frequency offset of a wireless communication channel based on a determined phase difference. In some embodiments, an access point is caused to receive one or more streams comprising one or more encoded long training field (LTF) symbols over the wireless communication channel from one or more user devices. The access point then determines a first phase of the wireless communication channel upon receipt of a first LTF symbol, and determines a second phase of the wireless communication channel upon receipt of a second LTF symbol. The access point next determines a phase difference between the first phase and the second phase. Based on the determined phase difference, the access point determines a carrier frequency offset of the wireless communication channel. Lastly, the access point modifies the wireless communication channel based at least in part on the determined CFO.

Abstract: People detection can provide valuable metrics that can be used by businesses, such as retail stores. Such information can be used to influence any number of business decisions such a employment hiring and product orders. The business value of this data hinges upon its accuracy. Thus, a method according to the principles of the current invention outputs metrics regarding people in a video frame within a stream of video frames through use of an object classifier configured to detect people. The method further comprises automatically updating the object classifier using data in at least a subset of the video frames in the stream of video frames.

Abstract: A method is directed to identifying spatially extensive image features that are not pixelwise contiguous in a two-dimensional digital image to be examined having pixels arranged in grid form in the form of a rectangular grid. Each pixel of the two-dimensional digital image is respectively assigned two coordinate values corresponding to their position in the rectangular grid of the digital image, and the respective coordinate values of adjoining pixels respectively differ by one pixel dimension. The individual pixels are run through in the order of a regular scan.

Abstract: The gesture-based searching of a line pattern representation amongst a collection of line pattern representations. Upon detecting an input gesture, a computing system matches the input gesture against each of multiple pattern representations. Each line pattern representation represents a line pattern having a changing value in a first dimension as a function of a value in a second dimension. At least some of the matched set may then be visualized to the user. The input gesture may be a literal line pattern to match against, or might be a gesture that has semantic meaning that describes search parameters of a line pattern to search for. The matched set may be presented so that a display parameter conveys a closeness of the match.

Abstract: A method, apparatus, system, and computer program product provide the ability to model a polyline boundary from point cloud data. Point cloud data is obtained and boundary cells are extracted. Potential boundary points are filtered from the boundary cells. Line segments are extracted from the potential boundary points and refined. A regularized polygon is obtained by intersecting the refined line segments.

Abstract: Methods and apparatuses for reducing the sensitivity of an optical signal to polarization. The method generally includes (i) reflecting the optical signal from a first mirror at a first angle relative to the optical signal to a second mirror at a second angle, and (ii) further reflecting the reflected optical signal from the second mirror to a receiver. The apparatus generally comprises (i) a first mirror at a first angle relative to an incident optical signal and configured to reflect the incident optical signal, (ii) a second mirror at a second angle configured to further reflect the reflected optical signal to a first receiver, and (iii) a lens configured to focus and/or collimate the optical signal or the reflected optical signal. The first angle is configured to reduce polarization of the reflected optical signal, thereby maximizing the intensity or power of the optical signal.

Abstract: A vehicle periphery monitoring apparatus detects the presence of a moving object along a vehicle periphery. The apparatus sets multiple detection lines along a horizontal axis of an image captured by a camera, and detects a brightness change of a pixel along the detection lines. With reference to the brightness change detected along the detection line and a parameter for determining whether such brightness change is caused by the moving object, the apparatus determines the presence of the moving object. In addition, the apparatus changes a determination condition for determining the moving object such that as the number of detection lines along which the brightness change is detected decreases, the harder it is to satisfy the determination condition for determining that the moving object is present.

Abstract: A computerized method for recognition of a logo is described herein. The method comprises obtaining a plurality of feed frame of a feed video wherein the feed video has a logo embedded therein. At least one feed frame from the plurality of feed frames is compared with each template from a plurality of templates. For each template compared with the feed frame, a correlation parameter is computed and the logo is recognized based on the computing.

Abstract: The present invention provides a method, system and/or a digital camera providing a geometrical transformation of deformed images of documents comprising text, by text line tracking, resulting in an image comprising parallel text lines. The transformed image is provided as an input to an OCR program either running in a computer system or in a processing element comprised in said digital camera.

Abstract: An efficient image retrieval method utilizing topological order distance (TOD) algorithm is disclosed. In one aspect, an image recognition method may include steps of receiving one or more input images; extracting one or more key features on the input images; determining an origin and one or more feature points on the key features on the input images; establishing a topological relationship in each dimension between the feature points on the key features; and utilizing the topological relationship in each dimension between the feature points to identify the input images in an image database. Comparing with the conventional image retrieval techniques, the TOD algorithm is advantageous because it is more efficient, scale and rotation invariant, and tolerant to affine and perspective transformation.

Abstract: Methods, apparatus, and articles of manufacture to measure geographical features using an image of a geographical location are disclosed. An example method includes dividing, with a processor, an image of a geographic area of interest into a plurality of geographical zones, the geographical zones being representative of different geographical areas having approximately equal physical areas, measuring, with the processor, a geographical feature represented in the image for corresponding ones of the plurality of geographical zones, storing descriptions for the geographical zones in a computer memory, and storing values representative of the geographical feature of the geographical zones.

Abstract: There are provided an environment recognition device and an environment recognition method. the environment recognition device retains beforehand shape information that is information on a shape of a specific object; obtains a luminance of each of target portions, formed by dividing a detection area, and extracting a target portion including an edge; obtains a relative distance of the target portion including an edge; and determines a specific object indicated with the shape information by performing a Hough transform on the target portion having the edge based on the shape information according to the relative distance.

Abstract: A document recognition system and method, where images are represented as a collection of primitive features whose spatial relations are represented as a graph. Useful subsets of all the possible subgraphs representing different portions of images are represented over a corpus of many images. The data structure is a lattice of subgraphs, and algorithms are provided means to build and use the graph lattice efficiently and effectively.

Abstract: Methods and apparatuses to reduce memory bandwidth consumption when executing a program that uses integral images are provided. A first integral image representation of a grayscale image may be computed using a first number of bits per pixel. A second integral image representation of the grayscale image may be computed using a second number of bits per pixel. Additionally, an image processing algorithm may be executed on the grayscale image, wherein the size of a rectangle to be used by the image processing algorithm can be determined. Based on the size of the rectangle to be used by the image processing algorithm, it may be determined whether to use the first integral image representation or second integral image representation for the image processing algorithm.

Abstract: Restoration of edges in rasterized, continuous tone images includes selection of a pair of neighboring pixels that exchange a restoration value to make a steeper transition from a low value on one side of a boundary between the two pixels to a high value at the other side thereof. Although any pair of neighboring pixels is selected, this operation is only performed when four pixels in a direction perpendicular to the boundary, of which the selected pair form the two center pixels, have a monotonely increasing value. The restoration value is based on a difference value of nearest neighboring pixels on the same side of the boundary. This avoids overshooting around edges in an image. The method is advantageous when resealing an image to a higher resolution which occurs when a low resolution image is embedded in a page that is to be printed with a high addressability printer system.

Abstract: A method for modeling an object, particularly suited to complex objects such as anatomical objects, and manipulating the modeled object in a CAD environment includes obtaining volumetric scan data of a region and segmenting the scan data to identify a first object to produce a first set of signed distance values on a grid. Wavelet analysis of the first set of signed distance values provides a function-based representation of the object. A signed distance value model of a second object is obtained, and one or both sets of signed distance values are manipulated to perform a CAD operation.

Abstract: An automated printout inspection system identifies glyphs in an image by calculating a connectedness score for each foreground pixel, and comparing this score with a specified threshold. The system further generates training images by simulating printouts from an impact printer, including the specifying of specific error types and their magnitudes. The simulated printouts are combined with scan images of real-world printout to train an automated printout inspection system. The inspection results of the automated system are compared with inspection results from human inspectors, and test parameters of the automated system are adjusted so that it renders inspection results within a specified range of the average human inspector.

Abstract: A novel phase-coded aperture, associated imaging system, and design method is disclosed. The optical imaging system includes a coded-aperture followed optically by a detector array and includes an image processor. A diffraction pattern in the form of a band-limited uniformly redundant array is formed on the detector array when focusable radiation from a point source in object space is modulation by the transmission function of the coded-aperture. Since diffraction effects cannot be ignored in the optical regime, an iterative phase retrieval method is used to calculate the phase-coded aperture transmission function. Correlation type processing can be applied for the image recovery.

Abstract: A method of and apparatus for grading a characteristic of an animal is described. The animal is guided to a detection area whereupon an image of the back of the animal is captured. Analysis of the image identifies anatomical points and determines angles at these points. The angles are then used to calculate a grading for a characteristic of the animal. An embodiment is presented for automating the determination of body score condition in dairy cows using seven angles determined at three anatomical points from an image over the back of the cow.

Abstract: We describe a method and apparatus for monitoring the dripping speed in IV process, as well as detecting the its end. The apparatus includes a camera and a processing unit to analyze the information from the acquired video. Features from the image sequence are extracted and the dripping speed is computed accordingly by discrete Fourier transform. The apparatus is also capable of detecting the end of the dripping process by finding the location of liquid surface in the drip chamber. We also describe the use of barcode to provide information to the monitoring device and program.

Abstract: A method for automatic mismatch correction is presented. The method includes identifying a feature of interest in a reference image volume and a target image volume. Furthermore, the method includes computing a cost matrix based on one or more pairs of image slices in the reference image volume and the target image volume. The method also includes identifying one or more longest common matching regions in the reference image volume and the target image volume based on the computed cost matrix. In addition, the method includes aligning the reference image volume and the target image volume based on the identified one or more longest common matching regions. A non-transitory computer readable medium including one or more tangible media, where the one or more tangible media include code adapted to perform the method for automatic mismatch correction is also presented.

Abstract: According to one aspect of the present invention, there is provided an image processing apparatus comprising: a thinning process unit which extracts a core line by applying a thinning process; a line width estimation unit which estimates an original line width of each pixel of the core line; a core line division unit which divides the core line; a monospaced line determination unit which determines, based on line width information, whether or not each of core lines divided by the core line division unit is a monospaced line; a connection relation information generation unit which generates connection relation information with another core line in association with each of the core lines; and a core line modification unit which modifies the core lines, based on pieces of line width information of core line pixels, a monospaced line determination result, and pieces of connection relation information generated.

Abstract: The present disclosure relates to a method for locating the iris in an image of an eye, comprising steps of locating the pupil in the image, of detecting positions of intensity steps of pixels located on a line passing through the pupil and transition zones between the iris and the cornea, on either side of the pupil, and of determining the center and the radius of a circle passing through the detected positions of the transitions.

Abstract: A method of estimating position and orientation of a vehicle using image data is provided. The method includes capturing an image of a region external to the vehicle using a camera mounted to the vehicle, and identifying in the image a set of feature points of the region. The method further includes subsequently capturing another image of the region from a different orientation of the camera, and identifying in the image the same set of feature points. A pose estimation of the vehicle is generated based upon the identified set of feature points and corresponding to the region. Each of the steps are repeated at with respect to a different region at least once so as to generate at least one succeeding pose estimation of the vehicle. The pose estimations are then propagated over a time interval by chaining the pose estimation and each succeeding pose estimation one with another according to a sequence in which each was generated.

Abstract: Embodiments disclosed include methods for connected component labeling including labeling groups of raw data as one or more regions, the labeling including designating one or more data structures as containing information about the one or more regions; designating one or more of the regions as one or more subregions to expose a spatial distribution of one or more region features; and arranging at least one memory array with a 1:1 correspondence to a data array associated with the raw data to enable one or more data structures to include feature labels of the one or more subregions, the 1:1 correspondence enabling acquisition of the one or more region features with a controllable precision.

Abstract: In an image data output processing apparatus of the present invention, an image matching section is capable of determining whether a similarity exists between each image of an N-up document and a reference document when input image data is indicative of the N-up document. An output process control section is capable of regulating an output process of each image in accordance with a result of determining whether the similarity exists between each image of the N-up document and the reference document. This allows detecting with high accuracy a document image under regulation on the output process and regulating the output process, when the input image data is indicative of an N-up document and includes the document image under regulation on the output process.

Abstract: A method for visual recognition of an object in an electronic image includes extracting unique points of an object to be learned and/or a target object. The unique points are obtained by cross-correlating the image with a structure. Generally, the structure and/or the size of the structure may vary to detect extremum information associated with the learned object and/or target object. An icon corresponding to each of the unique points is extracted. The size of the icon corresponds to the scale of the unique point. After extraction of the various icons, an object becomes a collection of icons. Each of these icons is un-rotated and normalized or resized to a constant size so it can be compared with other icons.

Abstract: A method for extracting a 3D terrain model for identifying at least buildings and terrain from LIDAR data is disclosed, comprising the steps of generating a point cloud representing terrain and buildings mapped by LIDAR; classifying points in the point cloud, the point cloud having ground and non-ground points, the non-ground points representing buildings and clutter; segmenting the non-ground points into buildings and clutter; and calculating a fit between at least one building segment and at least one rectilinear structure, wherein the fit yields the rectilinear structure with the fewest number of vertices.

Abstract: A method for changing the depth at which an image is focused is provided in a portable imaging device that includes an image recording arrangement that has a lens system, an image sensor for recording images, and a display arrangement for presentation of the recorded images. The method includes moving a movable focusing frame covering a part of a recorded image presented on the display to identify an area of interest in the presented image. The method also includes analyzing the image within the frame to obtain a setting required to focus the image within the frame and applying the obtained setting to the full presented image to focus the presented image at the depth defined by the current position of the movable focusing frame.

Abstract: A method and electronics circuit for processing very high resolution images or very high frame rate images in real time. Each pixel within a frame of pixels is compared to the neighboring pixels within the frame to determine if the pixel is part of a blob group. If the pixel is part of the blob group, the characteristics of the pixel are added to the statistics for the blob group. When a pixel overlaps two target blob groups, the two blob groups are combined to form one blob group. When the end of the frame is reached information about the blob groups in the frame is made available.

Abstract: A method for visual recognition of an object in an electronic image includes extracting unique points of an object to be learned and/or a target object. The unique points are obtained by cross-correlating the image with a structure. Generally, the structure and/or the size of the structure may vary to detect extremum information associated with the learned object and/or target object. An icon corresponding to each of the unique points is extracted. The size of the icon corresponds to the scale of the unique point. After extraction of the various icons, an object becomes a collection of icons. Each of these icons is un-rotated and normalized or resized to a constant size so it can be compared with other icons. One of the unique properties of these icons is their stability over scale and angle. Thus, this invention allows the recognition of an image(s) or object(s) from large number of trained images or objects very quickly.

Abstract: A system and method for determining inflection points in an image of an object includes obtaining the image of the object, performing binary image processing on a border of the image to obtain border points, selecting a predetermined number of the border points to fit a straight line, calculating a vertical distance between each selected border point and the straight line, and obtaining a total distance. The method further includes adding a new border point to the selected border points if the total distance is less than a predetermined value, so as to fit a new straight line and do a loop cycle, otherwise, regarding a last border point of the selected border points as an inflection point, and sequentially selecting the predetermined number of other border points to fit another new straight line.

Abstract: A method and electronics circuit for processing very high resolution images or very high frame rate images in real time. Each pixel within a frame of pixels is compared to the neighboring pixels within the frame to determine if the pixel is part of a blob group. If the pixel is part of the blob group, the characteristics of the pixel are added to the statistics for the blob group. When a pixel overlaps two target blob groups, the two blob groups are combined to form one blob group. When the end of the frame is reached information about the blob groups in the frame is made available.

Abstract: A method and system analyzes data acquired by image systems to more rapidly identify objects of interest in the data. In one embodiment, z-depth data are segmented such that neighboring image pixels having similar z-depths are given a common label. Blobs, or groups of pixels with a same label, may be defined to correspond to different objects. Blobs preferably are modeled as primitives to more rapidly identify objects in the acquired image. In some embodiments, a modified connected component analysis is carried out where image pixels are pre-grouped into regions of different depth values preferably using a depth value histogram. The histogram is divided into regions and image cluster centers are determined. A depth group value image containing blobs is obtained, with each pixel being assigned to one of the depth groups.

Abstract: Methods are constructed for defining equivalence relations on embedded manifolds. A continuous path in the space of equivalence relations that spans from “homologous” to “diffeomorphic” and another similar path with endpoints “homologous” and “isometric” are given. Underlying the methodology is a convolution algebra of homology sampling functions. The methodology has applications for describing geometrical arrangements in material science, molecular biology, data science and physics. The methods are implemented in program code stored in a computer readable media and executable on a computer system to provide data analysis functions for a user.

Abstract: A geospatial modeling system may include at least one geospatial information database to store stereo-geographic image data and geographic feature data. A processor may cooperate with the geospatial information database for generating cost coefficients defining a three-dimensional (3D) cost cube using image matching operators based upon the stereo-geographic image data, adjusting the cost coefficients of the 3D cost cube based upon the geographic feature data to generate an adjusted 3D cost cube, and generating a geospatial model based upon solving the adjusted 3D cost cube, e.g. for a best cost surface. The system and method provide an integrated approach to creating a geospatial model using available data from multiple sources.