Abstract: The present invention recites a method and computer program product for identifying one or more new pattern classes and incorporating the classes into a pattern recognition classifier as output classes. 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 then grouped into clusters according to the similarities between the pattern samples. Clusters that represent new pattern classes are identified via independent review. The classifier is then retrained to include the new pattern classes as output classes.

Abstract: A progressive self-learning (PSL) method is provided for enhancing wafer or mask defect inspection review and classification by identifying a plurality of wafer or mask defects, and by classifying each of the plurality of defects according to an extent of resemblance of each defect. The method having the steps of: performing image processing on a scanned defect image; aligning the scanned defect image with a just-stored digitized defect image; matching the scanned defect image with a just-stored digitized defect image; and classifying the scanned defect image.

Abstract: A system and method for automated print identification is described. In particular, the method includes providing an active search record, arranging the prints of the active search record according to print quality to generate a filtered active file database, providing the filtered active file database as the input to a match algorithm, detecting a print correlation corresponding to a print contained in the active file database, determining if a print match has been made, determining the strength of a match if a print match has been made, and determining the active search file record to be added into the active database or the matched active file records to be updated.

Abstract: A handwriting recognition device using fuzzy logic and cellular neural network for unconstrained handwritten numeral classification is provided. The current mode VLSI classifier has a I/O circuit for inputting and outputting a plurality of membership functions. An extraction unit comprising a CCD extractor with a CNN structure and a compression unit receives a to-be-recognized character having a plurality of input features for generating a plurality of features values after compression. A membership function generator stores the plurality of membership functions and receives the plurality of features values to generate a plurality of current-type membership degrees. A plurality of switched-current integrators receives the plurality of current-type membership degrees for generating a plurality of synthesis membership degrees.

Abstract: An electronic device stores a plurality of user's fingerprints in such a manner that each of a plurality of functions is stored in correspondence to any one of the fingerprints. Every time before using a function, the electronic device recognizes a user using a fingerprint corresponding to the function. For example, when the right thumb, right index finger and right middle finger are respectively stored in correspondence to an individual recognition function, imaging execution function and copy quantity setting function, at least three fingerprints are checked while the user is recognized by the electronic device, the copy quantity is set and imaging is accomplished. In such a manner that a user is multiply identified by a plurality of fingerprints during using the electronic device, the accuracy in recognizing a user and security can be enhanced with the structure as simple as the conventional one.

Abstract: Techniques for generating a fully-textured 3D model of a human face are disclosed. According to one aspect of the present invention, a fully automatic system is configured to generate a 3-D model from a sequence of images that are produced from a single camera when a user is moving his/her head steadily from one direction to another, for example, from the left side of the camera to the right side of the camera. The images are processed in a two-stage process. First, the camera pose information is estimated for each of the valid image frames. Second, a mesh model is generated with respect to the camera pose information, and subsequently is textured with reference to one or more of the image frames.

Abstract: A method and computer program product are disclosed for recognizing italic text in an optical character recognition system. A plurality of digital images of alphanumeric characters are created from a block of text, each comprising a plurality of rows of digital pixels. The digital images are preprocessed such that each image is normalized to equal size and the pixels within each image have a first value, such as “black,” or a second value, such as “white.” The position of the left-most pixel with a first value in each row is determined for each image. The position of each left-most pixel with a first value is recorded as an ordered pair including the row number and ordinal position within the row. A best-fit line and a corresponding slope are calculated for the ordered pairs via linear regression. The calculated slope is then compared to a predetermined threshold.

Abstract: In a fingerprint image, this fingerprint matching method extracts as skeleton stability the area of a zone which is drawn around a pre-determined attention point and which does not contain any minutiae that are bifurcations or endpoints in ridges forming a fingerprint pattern, and calculates similarity by collating said skeleton stability thus extracted with skeleton stability already registered.

Abstract: A method for extracting and matching gesture features of image is disclosed. An input gesture image is captured, and then a closed curve formed by a binary contour image of the gesture image is determined by preprocessing the gesture image. A curvature scale space (CSS) image of the gesture image is drawn based on the closed curve. Feature parameters of a plurality of sets of the gesture image are determined by extracting first plural peaks from the CSS image as basis points, and each feature parameter of the plurality of sets of the gesture image is compared with each feature parameter of a plurality of reference gesture shapes represented as a basis point of the maximal peak, thereby determining a gesture shape corresponding to the gesture image.

Abstract: An image processor for generating, from a surface image of an object having a pattern, a processed image having a distinct contrast between the pattern and a background, is provided. This image processor comprises a first smoothing filter for generating a first smoothed image; a second smoothing filter for generating a second smoothed image; a weighting unit for multiplying each of pixel values of the first smoothed image by a weight coefficient of 1 or more to determine a weighted pixel value; a differential image generator for subtracting each of weighted pixel values provided by the weighting unit from a pixel value of a corresponding pixel of the second smoothed image to generate a differential image; and an image extracting unit for extracting pixel values having the positive sign from the differential image to obtain the processed image. For example, this image processor can be utilized in a pattern recognition apparatus.