Abstract: An optical pattern recognition apparatus has an optical coordinate converter unit which simultaneously and in parallel converts an entered reference image and an entered input image from an entered coordinate system to a desired coordinate system to produce a coordinate-converted reference image and a coordinate-converted input image. A joint transform correlator receives the coordinate-converted reference image and the coordinate-converted input image and generates a correlation coefficient between the coordinate-converted reference image and the coordinate-converted input image so that a correlation between each entered reference and each entered input image is obtained to effect pattern recognition in real time. A feedback circuit normalizes the generated correlation coefficient between the coordinate-converted reference image and the coordinate-converted input image with a maximum correlation coefficient.

Abstract: An image decoding and recognition system and method comprising a fast heuristic algorithm using hidden Markov models (HMM). The new search algorithm, called an "iterative complete path" (ICP) algorithm, patterned after well-known branch-and-bound (B&B) methods, significantly reduces the complexity and improves the speed of HMM image decoding without sacrificing the optimality of the straightforward procedure. An advantageous form of the heuristic functions which is useful in applying the ICP algorithm to text-like images is described. The ICP algorithm is directly applicable to the separable type of finite-state source models. Also disclosed is a technique for transforming more general source models into such a separable form.

Abstract: In a mail sorting system 10, output pixel intensities of an optical scanner 18 have their contrast locally enhanced. A contrast enhancer 24 uses statistical methods (averaging, standard deviation) coupled with empirical stretch and offset data stored in a ROM 25, to enhance pixel contrast in a pipelined processing operation.

Abstract: An apparatus, and corresponding method for finding an area of interest in an input image, filters the input image with at least one compression filter to generate a final compressed image. The apparatus determines the location of an artifact in the final compressed image and then determines the location of the area of interest in the input image according to the location of the artifact in the final compressed image. An apparatus, and corresponding method for finding one or more artifacts in a two-dimensional image, receives a first row of the image and generates a list of regions in accordance with the first row. The apparatus receives a next row of the image and updates the list of regions in accordance with the next row. The apparatus determines whether a region in the list of regions corresponds to an artifact in the image and, if so, selects the region as the artifact.

Abstract: A top-down technique for character text recognition of an image comprises a left-to-right analysis of each image line. A current image portion is selected. Possible text prefixes are selected from a dictionary. The upper and lower text contours of the text prefixes are compared with a bitmap of the current image portion. A distance value is generated, indicating the quality of the comparison. The prefixes are then added to an agenda of prefixes. Based on the distance value, corresponding to the similarity of the upper shapes and lower shapes of the possible prefix to the bitmap of the image portion, a list of the text prefixes generating the best distance values is selected from the agenda. From the selected list, a new list of extended text prefixes is obtained from the dictionary and added to the agenda. The process is repeated until the current image portion ends.

Abstract: Pattern recognition system which provides an indication of the confidence with which a candidate is selected for an unknown pattern. The pattern recognition apparatus includes an image data input device, a host for segmenting the image data into unknown patterns, and a character recognition device for providing a candidate for each unknown pattern. The character recognition device includes a confidence level indicator for providing a confidence level indication. In one aspect, the confidence level indication is determined based on the proximity of an unknown pattern to a known pattern. In another aspect, the confidence level indication is determined based on the consistency with which the unknown pattern is recognized using different recognition functions. In yet another aspect, the confidence level indication is determined by ensuring that a candidate is not provided unless the candidate is closer than a predetermined distance from a known pattern.

Abstract: A system for recognizing handwritten words of cursive script by extracting temporal information from static text is described. The temporal information consists of the shape, location and temporal sequence of the segments, or "primitives", constituting the word and is translated from bit maps images of the words into digital information format. By comparing the digital information with formatted digital information for a plurality of known words, as may be stored in computer memory, a list of words having a relatively high probability of matching the static text is generated, and used for word recognition by a process of selecting the most likely match.

Abstract: Disclosed is a new self-reference signal processing technique for detecting the location of any nonregularities and defects in a periodic two-dimensional signal or image. Using high-resolution spectral estimation algorithms, the proposed technique first extracts the period and structure of repeated patterns from the image to sub-pixel resolution in both directions, and then produces a defect-free reference image for making comparison with the actual image. Since the technique acquires all its needed information from a single image, on the contrary to the existing methods, there is no need for a database image, a scaling procedure, or any apriori knowledge about the repetition period of the patterns.

Abstract: A method and apparatus are specialized in utilizing a two dimensional histogram for designating a ROI including accurate functional information of a lesion in tissues. A pair of medical images imaged at an identical position of a patient under different imaging conditions is provided by an image providing element. Each of the images extends two-dimensionally and consists of a plurality of pixels. A two-dimensional histogram is calculated by a histogram calculating element and displayed by a histogram displaying element. The histogram projects a function between pixel intensities of the images and frequencies of the pixel intensities. A desired region is then designated by a region designating element on the histogram displayed. Finally, pixels of the plural images equal in pixel intensities to pixels composing the desired designated region are specified by a pixel specifying element as a finally designated ROI.

Abstract: An automated truss inspection system for inspecting metal plate connected trusses includes a base structure having a truss feed end and a truss exit end, a truss presence detector and a bar code scanner. A truss structure sensor array and at least one metal sensor array are situated intermediate the truss feed end and truss exit end for detection of truss structure and absence thereof and detection of the presence/absence of metal, respectively. Finally, a rotary pulse generator is included for determining truss speed and distance traveled. All of the scanning and measuring devices are operatively connected to an analyzing and comparing device such as a computer which analyzes data received from those devices and constructs an X-Y coordinate outline image of the truss under inspection, the image then being compared to the ideal model of the truss under inspection to determine acceptability of the truss.

Abstract: A general approach is provided for the combined use of several sources of information in the automatic recognition of a consistent message. For each message unit (e.g., word) the total likelihood score is assumed to be the weighted sum of the likelihood scores resulting from the separate evaluation of each information source. Emphasis is placed on the estimation of weighing factors used in forming this total likelihood. This method can be applied, for example, to the decoding of a consistent message using both handwriting and speech recognition. The present invention includes three procedures which provide the optimal weighing coefficients.

Abstract: A dividing step (a) divides into segments at least one character line forming input characters to be read. A partial-figure forming step (b) forms partial figures by combining segments from among the thus obtained segments. A character reading step (c) attempts to read each of said partial figures as a character. A network forming step (d) forms a network wherein partial figures, among the partial figures, which have been read in said character-reading step (c) are used as nodes and said nodes are connected with one another by links, and wherein said links respectively have appropriate weights. An optimum-path selecting step (e) selects an optimum path from among paths existing in said network so that said optimum path comprises nodes, from among said nodes, which respectively correspond to said input characters.

Abstract: Any of multiple landmark regions of a scene being viewed by an image sensor are used for interring the location of a target region of the scene in which a pattern is to be inserted in order to avoid the problem of the view of a region being occluded. Because of changes, such as the pointing angle of the sensor, there is a slight difference in size and position in the location of a target region inferred from one landmark region and from another landmark region. this results in a jitter when a first landmark region being used for inferring the location of a target region is replaced by a second landmark region in response to the first landmark region being occluded. This jitter is eliminated by deriving an error-correcting component that compensates for the aforesaid slight difference in size and position in the location of a target region inferred from different landmark regions.

Abstract: A magnetic ink character recognition (MICR) system includes power management features so as to permit the unit to be powered by the auxilliary power provided by a generic host (e.g., an electronic cash register). The system also includes motorized conveyance features in combination with a track facilitating manual conveyance.

Abstract: An adaptive spatial filter for digital image processing selectively performs spatial high pass filtering of an input digital signal to produce an output pixel value having an enhanced signal level using a gain factor that effects a variable degree of enhancement of the input pixel, so that pixels whose activity values are close to an iteratively adjustable activity threshold level will be emphasized significantly less than pixels whose associated activity values are substantially above the threshold level. The activity threshold level is selectively updated at the end of a scan line in accordance with prescribed spatial activity criteria, so as to effectively prevent discontinuities in the threshold level during line-by-line adjustment and to take into account how `noisy` the background is.

Abstract: Disclosed is a system and a method for compressing digitized color video data by generating codes into a pattern of tables for regions in frames of the video data exhibiting certain patterns. An image frame in a video data stream has a plurality of pixels assigned digitized color and luminance data. The image frame is divided into a plurality of non-overlapping elementary units, with each elementary unit comprising a plurality of pixels. For elementary units exhibiting change from the prior frame in time, elementary units having pixels with differing digitized color and luminance data are selected for pattern matching. Each pixel of an elementary unit selected for pattern matching is mapped to a pattern value. A pattern value is the same for all pixels sharing the same color and luminance data in an elementary unit. The pattern values are then grouped based on relative position in the elementary unit into offsets into associated sets of pattern match tables and error condition tables.

Abstract: Data compression apparatus is provided for compressing an input stream of n-bit data samples, possible input sample values being notionally divided into groups each containing a number of sample values equal to a respective power of two.

Abstract: A wiring pattern inspection apparatus for inspecting an abnormality of a wiring pattern formed on a printed circuit board, which is equipped with an optical image pickup device for optically illuminating a surface of the printed circuit board including the wiring pattern to photoelectrically convert optical information of the printed circuit board surface due to the optical illumination into a grey level image. This grey level image is converted into a bi-level image which separates the grey level image into the wiring pattern side and a background side of the wiring pattern. Thereafter, the bi-level image is once contracted by a first size and then expanded by a second size so as to eliminate a micro conductive portion left on the printed circuit board or a micro pinhole which can be disregarded in the abnormality inspection, thereby preventing the excessive detection.

Abstract: An improved method is provided for extracting the edge boundary of an object presented as an ultrasound video image, in particular the edge boundary of the LV of the heart, which includes preprocessing algorithms, contour extraction algorithms and knowledge based contour following algorithms.

Abstract: An interactive image data transmission system uses the expertise of an image end user distant from the image collection point. An image with a wide angle field of view is transmitted to the end user in lossy-compressed form. The image is displayed at the receiving end with errors introduced by the lossy compression. The end user can select and specify a narrowed field of view, and the image capturing parameters are modified as necessary under end user control to capture a desired image of the narrow view field. The transmitting station transmits a lossy compressed version of the narrow view image which is displayed for the user, and the user selects at least one window of definite interest within the image. Then, additional image data sufficient to provide a correct display within the window(s) of interest is transmitted to the receiving station and the display is updated within the window(s). The additional data may be difference data in a variable length diagonal code.