Abstract: A system augments stylus keyboarding with shorthand gesturing. The system defines a shorthand symbol for each word according to its movement pattern on an optimized stylus keyboard. The system recognizes word patterns by identifying an input as a stroke, and then matching the stroke to a stored list of word patterns. The system then generates and displays the matched word to the user.

Abstract: Document processing and understanding can be important for a variety of applications such as office automation, creation of electronic manuals, online documentation and annotation, etc. The first step towards this process can involve the decomposition of a page into separate units describing the text, graphics, pictures etc. Towards that end, a method of page decomposition comprises segmenting an image into a plurality of blocks according to an amount of information contained in each block, classifying each block according to a corresponding frequency spectrum, and refining a classification of each block.

Abstract: An apparatus configured to discriminate the genuineness of a certified paper by exposing the paper to a light source and transporting the paper through an environment of an alternating-current electromagnetic field. Light detectors detect the light reflected from the paper and the light emitted by a fluorescent ink on the paper and output a waveform corresponding to the intensity of the detected light, reflected light, and the detected emitted light as the paper is scanned. The resulting waveforms are used to discriminate the genuineness of the paper. The fluorescent light and reflected light can be optically split and directed to separate detecting devices. Furthermore, the alternating-current electromagnetic field can be controlled by a voltage controller and configured to be driven only when the sensor detects light from the fluorescent ink.

Abstract: A method of applying a tone scale function that is compressive, expansive or a combination of both to a digital image, the method including decomposing the tone scale function into two or more composite functions that can be applied sequentially to the digital image; applying the first composite function to the digital image with a tone scale applicator to produce a tone scaled digital image; and applying the second composite function to the tone scaled digital image to produced an enhanced digital image.

Abstract: A method for determining dental alignment of a 3-dimensional model of one or more teeth of a patient comprises the steps of: (a) obtaining a radiograph of the teeth of the patient; (b) obtaining a digital image from the radiograph indicative of the dental alignment of the teeth relative to a dental arch of the patient; (c) overlaying the 3-dimensional model of the teeth with the digital image obtained from the radiograph; (d) determining vertical and horizontal mis-alignment of the teeth in the 3-dimensional model relative to the digital image obtained from the radiograph; and (e) adjusting the 3-dimensional model to correct for the mis-alignment, thereby producing an adjusted 3-dimensional model of the teeth that is corrected for the vertical and horizontal alignment of the teeth relative to the dental arch.

Abstract: A device for carrying out of biometric identification by comparing face shapes. The shape of a first face is compared with the shape of a second face. A map can be determined between the shapes. The shape, and hence the comparison, is substantially independent of pose and lighting, and also parts of the shape can be removed from the comparison in order to enhance the recognition.

Abstract: An apparatus for discriminating the genuineness of a security mark is provided having a first light detecting element for detecting a light emitted from a luminescent ink of the security mark placed in a specified environment and a second light detecting element for detecting a reflected light from a non-luminescent ink of the security mark. The first and second light detecting elements detect the light from the same location of a bank note. A third light detecting element detects the light intensity of a light emitted from a monitor mark that can be compared with a preset reference light intensity to calculate a deviation from the reference light intensity. The output of the first light detecting element can be converted into a corrected output value corresponding to the calculated deviation. The genuineness of the bank note can be discriminated based on the corrected output value.

Abstract: An automated symbolic recognition system and method includes multiple representations of each alphanumeric in the database by identifying a finite number of exemplars per alphanumeric symbol.

Abstract: A system and method are disclosed for effectively utilizing universal feature detectors to generate closely-correlated universal feature information for use in various content-retrieval procedures. In certain embodiments, the universal feature detectors may generate a multi-scale gaussian pyramid based upon selected original image data by iteratively performing a series of summing transforms. The universal feature detectors may also calculate multi-scale histograms corresponding to the multi-scale gaussian pyramid, and may then perform Radon transforms upon the multi-scale histograms to produce and store corresponding Radon transform lines. The universal feature detectors may also generate a multi-scale laplacian pyramid from the multi-scale gaussian pyramid by iteratively performing a series of difference transforms. The resultant universal feature information may then advantageously provide various types of required feature information with optimal binding characteristics.

Abstract: A method of verifying identity which includes recording coordinates of a reference signature which include two dimensional space coordinates x and y exerted by a writer's writing instrument on a recording medium. The method measures and records an indicium, which identifies a reference biometric within the reference signature having a selected characteristic. Next the reference biometric and the indicium are placed on a portable, readable substrate. The coordinates of an unknown signature which include two dimensional space coordinates x and y covered by a writer's writing instrument are also recorded. The method further includes reading and storing the indicium and the reference biometric in memory and using the indicium to locate an extracted biometric within the unknown signature. The extracted biometric is compared to the reference biometric to determine if they match within predetermined threshold criteria.

Abstract: An automated symbolic recognition system and method includes an evidence-based technique to significantly reduce the alphanumeric candidate symbol list.

Abstract: A method and apparatus for classifying a plurality of elements in images, where electronic images of a field of view containing elements are formed. Each of the elements has a plurality of features. A first subgroup of the plurality of features from the images of the plurality of elements are extracted and processed to segregate the plurality of elements into first and second groups. A classification class only for each of the elements in the first group is determined by selecting and processing a second subgroup of the extracted features to determine a physical characteristic of the element, and by selecting and processing a third subgroup of the extracted features in response to the determined physical characteristic to determine a classification class of the element. The second group of elements bypasses the determination of classification class.

Abstract: An automated symbolic recognition system and method includes the basis for the system's feature set by identifying (a) a set of logical symbols comprising a finite class of arcpolys (lines and arcs and a point) that to the exclusion of the point, each member class has a unique (distinct) orientation, and (b) a set of subclass symbols per logical class of symbol representing a finite subclass of arcpolys (lines and arcs and a point) that to the exclusion of the point, each subclass member has a unique (distinct) extreme points size and/or depth size.

Abstract: A method for human face detection that detects faces independently of their particular poses and simultaneously estimates those poses. Our method exhibits an immunity to variations in skin color, eyeglasses, facial hair, lighting, scale and facial expressions, and others. In operation, we train a convolutional neural network to map face images to points on a face manifold, and non-face images to points far away from that manifold, wherein that manifold is parameterized by facial pose. Conceptually, we view a pose parameter as a latent variable, which may be inferred through an energy-minimization process. To train systems based upon our inventive method, we derive a new type of discriminative loss function that is tailored to such detection tasks. Our method enables a multi-view detector that can detect faces in a variety of poses, for example, looking left or right (yaw axis), up or down (pitch axis), or tilting left or right (roll axis).

Abstract: Disclosed is a diagnosis supporting system that acquires a reference image signal of a subject that is illuminated with visible light and a fluorescent image signal of the subject that is excited by irradiation with excitation light, calculates a difference between brightness level of a pixel in the reference image signal and brightness level of a pixel in the fluorescent image signal at the corresponding position for all of target pixels and generates an observation image signal by converting the pixels in the reference image signal whose differences are larger than a predetermined value into red pixels. In the case of high contrast, all of the pixels become target pixels to be processed. In the case of low contrast, the number of the target pixels is reduced to ¼ of all the pixels.

Abstract: In the neighborhood of each connecting region of a panel substrate, tape automated bonding (TAB) substrates, and a flexible substrate, marks for a visual inspection are provided at a preset distance apart from the connecting region. According to the present marks, inferior products are eliminated with a high rate of accuracy upon visual inspection after completing a connection of display panel.

Abstract: Apparatus and methods for converting raster illustrated parts images into intelligent vector-layered files. The method involves recognizing and removing reference labels from the raster illustrated parts image to produce a reference label scrubbed file. Reference lines are recognized and removed from the reference label scrubbed file to produce a scrubbed file. The scrubbed file includes a reusable base graphic. The scrubbed file is converted to a vector file in which the reusable base graphic is embedded. One or more vector layers are added to the vector file to produce the intelligent vector-layered file. Each vector layer includes vector elements corresponding to one of the recognized reference labels and its one or more reference lines.

Abstract: A method and computer program product are disclosed for identifying output classes with multi-modal dispersion in feature space and incorporating multi-modal structure into a pattern recognition system architecture. A plurality of input patterns, determined not to be associated with any of a set of at least one represented output class by a pattern recognition classifier, are rejected. The rejected pattern samples are grouped into clusters according to the similarities between the pattern samples. Clusters that contain samples associated with a represented output class are identified via independent review. The classifier is then retrained to recognize the identified clusters as output pseudoclasses separate from the represented output class with which they are associated. The system architecture is reorganized to incorporate the output pseudoclasses. The output pseudoclasses are rejoined to their associated class after classification.

Abstract: A knowledge-based hierarchical method for detecting regions of interests (ROIs) uses prior knowledge of the targets and the image resolution in detecting ROIs. The result produces ROIs that contain only one target that is completely enclosed within the ROI. The detected ROI can conform to the shape of the target even if the target is of irregular shape. Furthermore, the method works well with images that contain connected targets or targets broken into pieces. The method is not sensitive to contrast levels and is robust to noise. Thus, this method effectively detects ROIs in common real world imagery that has a low resolution without costly processing while providing fast and robust results.

Abstract: A chromaticity diagram is divided into a plurality of sections by a primary color line and a complementary color line of a color that does not cause hue turn, then it is decided to which section a color signal being output from an imaging device belongs, and then a color correcting operation is performed for the color signal every section. Accordingly, it is not needed to detect the flesh color itself when the hue turn of the flesh color is corrected by the color correcting operation. Any flesh color can be decided easily according to whether or not the color signal being output from the imaging device has entered into the section including the flesh color. Therefore, the appropriate color correcting operation can be applied to the flesh color.