Abstract: A circuit for performing neural network computations for a neural network comprising a plurality of layers, the circuit comprising: activation circuitry configured to receive a vector of accumulated values and configured to apply a function to each accumulated value to generate a vector of activation values; and normalization circuitry coupled to the activation circuitry and configured to generate a respective normalized value from each activation value.

Abstract: In a method and apparatus for optically inspecting an FPD in a reduced time, the apparatus for inspecting the flat panel display (FPD) comprises a lighting device separated from a top surface of an object and emitting light toward the top surface of the object, a photographing device separated from the top surface of the object and photographing an image using light emitted from the lighting device and then reflected by the top surface of the object, a controller controlling the lighting device and the photographing device to move in a plane separated from and parallel to the top surface of the object, and an image processor connected to the photographing device and processing the photographed image, wherein the image processor samples some of pixels included in the photographed image and converts the sampled pixels to binary form.

Abstract: A system and method for generating a direction confidence measure includes a camera sensor device that captures blur images of a photographic target. A depth estimator calculates matching errors for the blur images. The depth estimator then generates the direction confidence measure by utilizing the matching errors and a dynamic optimization constant that is selected depending upon image characteristics of the blur images.

Abstract: The user interaction system comprises a portable pointing device (101) connected to a camera (102) and sending pictures to a digital signal processor (120), capable of recognizing an object (130) and a command given by the user (100) by moving the pointing device (101) in a specific way, and controlling an electrical apparatus (110) on the basis of this recognition.

Abstract: A data processing apparatus that executes determining processing, using a plurality of stages, for determining whether or not a partial image sequentially extracted from an image of each frame of a moving image corresponds to a specific pattern, assigns a plurality of discriminators to each stage such that a plurality of partial images are processed in parallel. The data processing apparatus divides an image into a plurality of regions, and, for the image of each region, calculates a passage rate or accumulated passage rate from a ratio between the number of partial images input to a stage and the number of partial images determined to correspond to the specific pattern. The assignment of the discriminators to each stage is changed based on the passage rate or accumulated passage rate of the image processed immediately of a region to which the partial image extracted from the image being processed belongs.

Abstract: In one respect, provided are systems, methods and techniques in which local regions within an image are processed to provide fuzzy classification scores, which are calculated by determining changes in pixel values along a number of different directions. The resulting fuzzy classification scores are then used to detect or identify edge-containing or texture-containing regions, or to otherwise process the image regions differentially according to their fuzzy classification score. In another respect, provided are systems, methods and techniques for differential processing of different areas in an image. The differential processing in this case is based on calculated measures of local activity, which indicate features in corresponding local regions, and also based on calculated measures of local pixel-value variations, which indicate an amount of variation in pixel values across the corresponding local regions.

Abstract: An apparatus and method for encoding and decoding a multi-view image including a stereoscopic image are provided. The apparatus for encoding a multi-view image includes a base layer encoding unit that encodes a base layer image to generate a base layer bit stream, a view-based conversion unit that performs view-based conversion of the base layer image to generate a view-converted base layer image, a subtractor obtaining a residual between a enhancement layer image and the view-converted base layer image, and an enhancement layer encoding unit that encodes the obtained residual to generate an enhancement layer bit stream.

Abstract: It is provided a computer system for matching between images/graphs comprising a processing unit, a storage unit, an input unit and a data recording unit, wherein the computer system includes a feature pattern generating unit for generating a set of feature patterns of the image and the graph from the received image and the received graph, segmenting each of the image and the graph into a plurality of regions centered on a sample point, and determining characteristics of the plurality of regions; a correspondence computing unit for computing correspondence between the image and the graph to be matched by comparing the generated feature patterns; a transform unit for transforming a coordinate system of at least one of the image and the graph to be matched; and a combination unit for combining the image and the graph that have been subjected to the transform of the coordinate system.

Abstract: An image processing apparatus includes: a data obtaining section for obtaining input image data; a memory in which reference image data or features of a reference image is stored; and a similarity determination process section for performing a determination process in which it is determined whether the input image data is image data corresponding to the reference image or not. The similarity determination process section changes the determination process in accordance with related information of the input image data. Consequently, it is possible to realize an image processing apparatus capable of determining a similarity between input image data and a reference image and restricting a process on the input image data in accordance with the result of the determination.

Abstract: A tag information production apparatus is disclosed which can apply tag information to a main image through a realistic amount of work. A time information acquisition section acquires time information of a main image selected on a working screen, and a key region setting section sets a key region at a position corresponding to a partial region in the main image specified arbitrarily by an operator. An identification information application section applies, to the set key region, region identification information which allows identification between the key region and a peripheral region. A key image production section produces a key image to which time information same as the time information of the main image is applied and in which the region identification information is applied to the key region set through the working screen.

Abstract: Disclosed herein is a production apparatus for index information with link information. The apparatus includes a database in which a corresponding relationship between region identification information applied to a key region which provides a partial region in a main image specified arbitrarily by an operator and link information for accessing to related information of the partial region is listed. The apparatus further includes an information embedding section configured to refer to said database to embed, into index information in which a corresponding relationship between the region identification information applied to the key region and time information is listed, the corresponding link information.

Abstract: A decoding apparatus includes a distribution generating section and a correcting section. The distribution generating section generates a frequency distribution of quantization index values. The correcting section corrects dequantization values corresponding to the quantization index values based on the frequency distribution of quantization index values generated by the distribution generating section.

Abstract: Visual tracking over a sequence of images is formulated by defining an object class and one or more background classes. The most discriminant features available in the images are then used to select a portion of each image as belonging to the object class. Fisher's linear discriminant method is used to project high-dimensional image data onto a lower-dimensional space, e.g., a line, and perform classification in the lower-dimensional space. The projection function is incrementally updated.

Abstract: The invention relates to a method for encoding a digital image, in which method the digital image is divided into blocks (C, L, U, UL, UR). In the method a spatial prediction for a block (C) is performed to reduce the amount of information to be transmitted, wherein at least one prediction method (P1-P13) is defined. In the method a classification is determined for at least one neighboring block (L, U) of said block (C) to be predicted according to the contents of the neighboring block (L, U), and a prediction method (P1-P13) is selected for the current block (C) on the basis of at least one classification.

Abstract: A system and method for estimating noise using measurement based parametric fitting non-uniformity correction is disclosed. Fixed pattern noise (“FPN”) is estimating from an overall noise component within a detection system to enhance candidate target detection and tracking. A sensor in the detection system receives energy, such as radiant flux, that is converted to a digital image. A non-uniformity correction device generates an estimated FPN according to an applicable temperate range and integration time. A memory storing an array of coefficients is accessed to determine the estimated FPN. The valves within the array of coefficients are based on actual FPN measurements that are parametrically fitted.

Abstract: A system is presented for recognizing and identifying postal indicia on an envelope. The system includes an image acquisition element that acquires a first image, representing a first side of the envelope, and a second image, representing a second side of the envelope, and generates first and second candidate images from respective opposing corners of the first image and third and fourth candidate images from respective opposing corners of the second image. A feature extractor that, for each candidate image, divides the candidate image into a plurality of regions, extracts a plurality of numerical feature values from each of the plurality of regions, and recombines the plurality of feature values into a feature vector that represents the image. A classification element classifies the image into one of a plurality of output classes representing various types of postal indicia according to the numerical feature vector.

Abstract: In one respect, provided are systems, methods and techniques in which local regions within an image are processed to provide fuzzy classification scores, which are calculated by determining changes in pixel values along a number of different directions. The resulting fuzzy classification scores are then used to detect or identify edge-containing or texture-containing regions, or to otherwise process the image regions differentially according to their fuzzy classification score. In another respect, provided are systems, methods and techniques for differential processing of different areas in an image. The differential processing in this case is based on calculated measures of local activity, which indicate features in corresponding local regions, and also based on calculated measures of local pixel-value variations, which indicate an amount of variation in pixel values across the corresponding local regions.

Abstract: A system for recognizing and identifying postal indicia on an envelope. This includes an image acquisition element that acquires a first image, representing a first side of the envelope, and a second image, representing a second side of the envelope. A feature extractor, for each of the first and second image, extracts a plurality of numerical feature values from each image as respective first and second feature vectors that represent the envelope. An orientation classification element classifies the envelope into one of a plurality of output classes representing a plurality of possible orientations according to the first and second feature vectors.

Abstract: A system is presented for recognizing and identifying postal indicia on an envelope. The system includes an image acquisition element that acquires at least one binarized image of an envelope and isolates at least one region of interest from the at least one image. A candidate locator attempts to locate at least one candidate object within the at least one region of interest according to the distribution of dark pixels within the at least one region of interest. A feature extractor extracts an associated set of numerical feature values from each of the at least one candidate object. A classification element classifies each candidate object according to its associated set of numerical feature values.

Abstract: Systems and methods are provided for determining the orientation of an envelope. A plurality of classification elements are each operative to analyze at least one image of the envelope and produce at least one output value indicative of the orientation of the envelope. An arbitrator determines an associated orientation for the envelope according to the plurality of output values provided by the plurality of classification elements.

Abstract: A photosensing circuit, system and method. The photosensing circuit includes a photodetector and a network coupled to the photodetector. The network includes a plurality of elements and at least a first and second output terminals, the elements of the network being selected such that a difference of signals flowing out of the first and second output terminals is directly proportional to an amount of light impinging on the photodetector and is substantially free of electrical noise.

Abstract: The present invention relates to an image processing apparatus, and more particularly, to a device and method capable of binarizing pixels by removing an isolated point. The pixel binarization method contemplates the steps of: establishing a local window of pixels including a central pixel and a plurality of neighboring pixels; analyzing the neighboring pixels within the local window and detecting a minimum brightness value and a maximum brightness value exhibited by the neighboring pixels; comparing a brightness value exhibited by the central pixel with the minimum brightness value; and determining the central pixel to be an isolated point and outputting the central pixel as a white pixel when the brightness value exhibited by the central pixel is less than the minimum brightness value.

Abstract: A nonlinear resistive network, which includes switches in the data paths, is provided for identifying, isolating, and/or rejecting outliers in a sensor image. During normal operation, images are received with all the switches in a closed (conducting) state. If the data at a given pixel is different from its immediate neighbors by a predetermined threshold voltage, its switch is opened. A readout of the state of all the switches in the network yields a map of points sources, and a readout of the network voltages yields a noise-free image. Because the threshold voltage can be controlled externally, various strategies may be implemented for identification of the outliers in a computer vision system.