Abstract: A system includes an imager that images an object, which includes a marker that is visually obscured, a marker identifier that identifies markers in the image of the object, and an image processor that combines the image of the object and a visual representation for the identified marker into a combined image, wherein a presence of a visual representation for the marker included with the object in the combined image indicates the object is authentic. A method includes imaging an object that includes a visually obscured marker and generating an image of the object, identifying the marker in the image of the object, generating a combined image that includes the image of the object and a visual representation of the identified marker, presenting the combined image, wherein the object is authenticated as not counterfeit in response to the visual representation of the marker corresponding to a reference authentication marker for the object.

Abstract: An “active learning” method trains a compact classifier for view-based object recognition. The method actively generates its own training data. Specifically, the generation of synthetic training images is controlled within an iterative training process. Valuable and/or informative object views are found in a low-dimensional rendering space and then added iteratively to the training set. In each iteration, new views are generated. A sparse training set is iteratively generated by searching for local minima of a classifier's output in a low-dimensional space of rendering parameters. An initial training set is generated. The classifier is trained using the training set. Local minima are found of the classifier's output in the low-dimensional rendering space. Images are rendered at the local minima. The newly-rendered images are added to the training set. The procedure is repeated so that the classifier is retrained using the modified training set.

Abstract: In one embodiment, a character recognition result verification apparatus has a group generation section and a verification image generation section. The group generation section generates a group including a plurality of character images recognized as the same character from a document image including a plurality of character images. The verification image generation section generates a verification image including a first region and a second region by superimposing the plurality of character images included in the generated group. The first region corresponds to a pixel having the same pixel value in all of the plurality of character images. The second region corresponds to a pixel having the same pixel value in a part of the plurality of character images.

Abstract: A method is disclosed for recognition of high-dimensional data in the presence of occlusion, including: receiving a target data that includes an occlusion and is of an unknown class, wherein the target data includes a known object; sampling a plurality of training data files comprising a plurality of distinct classes of the same object as that of the target data; and identifying the class of the target data through linear superposition of the sampled training data files using l1 minimization, wherein a linear superposition with a sparsest number of coefficients is used to identify the class of the target data.

Abstract: As to each section making up of an original image, a CPU of an image processing apparatus calculates a distance from a predetermined position, for example, a center of the original image to a center of the section, sets a probability in accordance with the calculated distance, determines whether or not the section is a section where an object is to be drawn according to the set probability, and draws an object in each section which is determined as a section to be drawn.

Abstract: A method for producing a slide show video from a collection of hardcopy media, the method includes digitizing the media and detecting handwritten information and estimating the age of the media; determining an order of presentation for the slide show video based on the detected handwritten information and estimated ages; and producing a slide show video from the hardcopy media using the determined order of presentation.

Abstract: In a liquid crystal array inspection that acquires an imaging picture by scanning a liquid crystal substrate in a two-dimensional manner with an electron beam to inspect a liquid crystal substrate array based on the imaging picture, the imaging picture obtained by imaging a stage with the electron beam is used to determine the amounts of displacement of an imaging range of each electron gun in an X direction and a Y direction. Amounts of correction for correcting displacements of the imaging range of each electron gun in the X direction and the Y direction are calculated according to the determined amounts of displacement. The displacement in the X direction is corrected by controlling the scanning with the electron beam in the X direction, and the displacement in the Y direction is corrected by aligning a mounting position of each electron gun in the Y direction.

Abstract: A CPU of an image processing apparatus detects an outline from an original image, selects from the positions in association with the detected outline a position to be drawn at a ratio in correspondence with a predetermined ratio, for example, the number of positions, and draws an object at the selected position.

Abstract: An image processing device includes a storage module, character recognition module, a circumscribed rectangle extraction module, a ratio extraction module, and a character size calculation module. The storage module stores a reference ratio between a reference size of a reference circumscribed rectangle and a reference character size in a reference character image representing a reference character in association with a reference character identification code which uniquely identified the reference character. The character recognition module recognizes a character image in an image to get a character identification code from the recognized character image. The circumscribed rectangle extraction module extracts a circumscribed rectangle of the character image. The ratio extraction module extracts the reference ratio corresponding to the reference character identification code stored in the storage module based on the character identification code.

Abstract: A digital signature system includes an Electronic Data Interchange (EDI) server and a portable terminal communicating with the EDI server. The portable terminal includes a touch-sensitive display, and a processing unit includes a receiving module configured for receiving a file from the EDI server, a signature mode control module configured for providing a signature file mode for users to select, a signature file creating module configured for recording signature input through the touch-sensitive display and creating a signature file to record the input signature if the signature file mode is selected, and a file combining module configured for combining the signature file and the file from the EDI server into a composite signature file. A related method is also provided.

Abstract: A handwriting compound system is provided. The system includes a handwriting recognition module that receives a proper number of letters from a user, detects a distance between phonemes/syllables, and shapes and locations of the phonemes/syllables, and recognizes the user's handwriting. The system also includes a point designating module, a reference point setting module and a handwriting transforming module. The point designating module designates a particular one of the subdivided positions in each phoneme recognized by the handwriting recognition module. The reference point setting module sets a reference point, serving as reference for the alteration of a phoneme/syllable, inside or outside the user's input phoneme/syllable recognized by the handwriting recognition module. The handwriting transforming module designates x and y coordinates of a first point with respect to the reference point.

Abstract: A system for hiding secret data in halftone images is disclosed. The system includes an image input module, a halftoning processing module and an image output module. The image input module reads an original image data and a secret image data. The halftoning processing module is coupled to the image input module and receives the original image data and the secret image data, performs a first error diffusion process to the original image to generate a first halftone image and performs a second error diffusion process to the original image data and the secret image data to generate a second halftone image. The image output image module is coupled to the halftoning processing module and outputs the first halftone image and the second halftone image, wherein the second error diffusion process comprises adding or subtracting a variable noise strength to or from each pixel of the secret image.

Abstract: Disclosed is a method of searching a digital image of a document for a predetermined keyword. The method identifies a word in the digital image, the word comprising one or more shapes. A test matrix comprising a difference vector for each character of the word is generated, and a template matrix comprising a difference vector for each shape of the keyword is also generated, wherein a difference vector represents the differences between the visual features of a respective shape and the visual features of a collection of reference shapes. A measure of similarity between the word and the keyword is generated by comparing the test matrix and the template matrix.

Abstract: Methods, systems , and apparatus, including computer programs encoded on a computer storage medium, are disclosed for reducing the impact of lighting conditions and biometric distortions, while providing a low-computation solution for reasonably effective (low threshold) face recognition. In one aspect, the methods include processing a captured image of a face of a user seeking to access a resource by conforming a subset of the captured face image to a reference model. The reference model corresponds to a high information portion of human faces. The methods further include comparing the processed captured image to at least one target profile corresponding to a user associated with the resource, and selectively recognizing the user seeking access to the resource based on a result of said comparing.

Abstract: A motion detecting includes a fetcher which repeatedly fetches an object scene image having a designated resolution. An assigner assigns a plurality of areas each of which has a representative point to the object scene image in a manner to have an overlapping amount different depending on a size of the designated resolution. A divider divides each of a plurality of images respectively corresponding to the plurality of areas assigned by the assigner, into a plurality of partial images, by using the representative points as a base point. A detector detects a difference in brightness between a pixel corresponding to the representative point and surrounding pixels, from each of the plurality of partial images divided by the divider. A creator creates motion information indicating a motion of the object scene image fetched by the fetcher, based on a detection result of the detector.