Abstract: A business form handling system includes an image reading device for reading a business form and a terminal apparatus installed at the window for discriminatively deciding whether a read image corresponds to a first species of business form by referencing form species identifying definition information stored previously in a file. When the read image corresponds to the first species of business form, the terminal apparatus cuts an image from a field corresponding to a format inherent to the first species of business form to thereby recognize contents of statement from the cut image and executes processing in accordance with result of the recognition. Unless the read image corresponds to the first species of business form, the terminal apparatus decides discriminatively whether the read image corresponds to a second species of business form by referencing the form species identifying definition information.

Abstract: Image partitioning methods, which may be implemented by software or hardware on a computer, partition a set of multiple overlapping images to be transmitted so that only the visible portions of the images are transmitted. The underlying partitioning algorithm uses rectangles to iteratively partition the image set.

Abstract: A document processing system including a plurality of model images stored in a memory is provided. The model images are represented by a first information set with the first information set varying as a function of object-based coordinates. At least one of the plurality of model images includes a text containing region having index information intended for use in storing one or more document pages. Moreover, a query image represented as a second set of information varying as a function of object-based coordinates is provided. In practice, an image localization module corresponds the second set of information with a portion of the first set of information to obtain the text containing region, and a text extraction module extracts the index information from the text containing region to facilitate the storing of the one or more document pages.

Abstract: Disclosed is a method for pre-selecting pixels in data structures which demonstrate significant intensity and/or color gradients with respect to surrounding pixels, (tyically the result of depicted-object-edge-effects in the data structure), in “X” and/or “Y” and/or “Z” direction(s), and utilizing the pre-selected pixels in data structure characterizing spatial first and second central moment calculations, to the end that data structure characterizing vector(s) are constructed.

Abstract: Methods and apparatus for managing a source image that is output in a rotated or skewed orientation to a raster buffer. The source image includes a plurality of source samples arranged in source rows having a source direction. The raster buffer is oriented in a destination space and receives the rotated or skewed image as a series of samples in rows defined by a destination direction. The raster buffer provides rows of samples to a rendering device for rendering scan lines for output on a raster device. The method includes determining an angle between the source direction and the destination direction and adjusting the organization of the source samples to align source samples in scan rows where each scan line can be rendered from the raster buffer using at most a predetermined small number of scan rows.

Abstract: There is provided an image processing method for performing magnification processing at a specified magnifying power. The method divides the magnification processing at the specified magnifying power into a first magnification processing step and a second magnification processing step and performs the first and second magnification processing steps. A start position of the first magnification processing step and a start position of the second magnification processing step are dislocated relatively to each other so that phases of pixel positions in the first and second magnification processing steps are reversed to each other. There is also provided an image processing apparatus for implementing the method. Artifacts that are generated in image magnification (enlargement and reduction) processing can be suppressed.

Abstract: For displaying a selected zone of images belonging to a sequence of video images, each in the form of a matrix of pixels, on a display unit, a signal is generated by a fixed large-field camera having a definition that is greater than that of the display unit. A TV user or an operator at a control station selects a zone of format corresponding to that of the display unit in a field of view of the camera. The pixels of a modified image having the size and definition of the display unit are generated by interpolation from the pixels in the selected zone of the image supplied by the large-field camera. A device for implementing the method includes a module for warping the selected zone by operations of interpolation and resampling.

Abstract: A data compression method and system that include the substitution of a substring of data characters located at a first position in a stream of data characters with a substitution code. The substitution code includes a reference to a previous position in the stream of data characters at which is located a substring of data characters that matches the substring of data characters which are being substituted located at the first position. The substitution code also includes an indication of the size of the substituted substring. The reference in the substitution code is a backwards offset to the previous position relative to the first position. According to a further aspect, Huffman encoding can be applied to the backward offsets, the substring lengths, the consecutive literal character lengths, and the literal characters themselves to reduce the data requirement size.

Abstract: A system and method for accurately mapping between image coordinates and geo-coordinates, called geo-spatial registration. The system utilizes the imagery and terrain information contained in the geo-spatial database to precisely align geodetically calibrated reference imagery with an input image, e.g., dynamically generated video images, and thus achieve a high accuracy identification of locations within the scene. When a sensor, such as a video camera, images a scene contained in the geo-spatial database, the system recalls a reference image pertaining to the imaged scene. This reference image is aligned very accurately with the sensor's images using a parametric transformation. Thereafter, other information that is associated with the reference image can easily be overlaid upon or otherwise associated with the sensor imagery.

Abstract: The scanned image registration systems and methods according to this invention register sets of scanned subimages of an original document into a composite scanned image of the original document. The subimages are formed by scanning portions of the original document on a scanner having a scan area, or window or footprint, that is smaller than the size of the document. An appropriate set of affine parameters, or more generally, transformation parameters, is obtained for each subimage relative to a fixed coordinate system. The set of affine parameters for each subimage defines the rotational and translational deformation necessary to place that subimage into the fixed coordinate system. The fixed coordinate system is preferably defined by selecting one of the subimages as a base image. The other images are then warped into registration with the selected base subimage based on the determined affine parameters.

Abstract: A topocompositional image of an inanimate object or surface can be generated that comprises data that relates to a topographical image, data that relates to a chemical compositional image, and combining such topographical and chemical compositional data to form a composite image. Conceptually a laser or other suitable imaging source that functions to induce both topographical and chemical compositional data directs a laser beam into a scanning optics source. The scanning optics source directs an active beam to a point on a surface. A sensor subsequently collects data and information from the surface by receiving the reflection or deflection of the active beams in order to provide position, chemical compositional and topographical data that represent the interaction of the active beam with a point. Position data, chemical compositional data and topographical data are transmitted along individual feeds to a data analysis component where they are analyzed and combined in order to generate a composite image.

Abstract: The disclosed system has particular applications to image archives and reediting. Given an image of an article having graphic objects including texts and pictures, the disclosed system first distinguishes graphic objects from the background by segmenting the image so as to produce an object image comprising a foreground, wherein the foreground corresponds to the graphic objects in the image. With respect to the foreground, the disclosed system “repaints” those spaces previously occupied by the graphic objects, referred to as object spaces, with the background color in the image as such the background continuity is preserved when the graphic objects are rendered and subsequently superimposed onto the repainted background. To account for various backgrounds, the disclosed system uses a point-wise patching approach followed by a one-dimensional patching approach if the object spaces contain two different colors.

Abstract: A system and method for accurately mapping between camera coordinates and geo-coordinates, called geo-spatial registration, using a Euclidean model. The system utilizes the imagery and terrain information contained in the geo-spatial database to precisely align geographically calibrated reference imagery with an input image, e.g., dynamically generated video images, and thus achieve a high accuracy identification of locations within the scene. When a sensor, such as a video camera, images a scene contained in the geo-spatial database, the system recalls a reference image pertaining to the imaged scene. This reference image is aligned very accurately with the sensor's images using a parametric transformation produced by a Euclidean model. Thereafter, other information that is associated with the reference image can easily be overlaid upon or otherwise associated with the sensor imagery.

Abstract: The invention provides a method for transforming pixels of a digital image by a preselected-transformation. The method includes acts for receiving pixel values of a selected pixel block from an original image provider; determining transformed pixel values for an image block of the selected pixel block under the transformation; and burst writing the transformed pixel values to a random access memory.

Abstract: An apparatus and method for generating automated multi-view three dimensional pose and geometry estimation for the insertion of realistic and authentic views of synthetic objects into a real scene. A multi-view three dimensional estimation routine comprising the steps of feature tracking, pairwise camera pose estimation, computing camera pose for overlapping sequences and performing a global block adjustment to provide camera pose and scene geometric information for each frame of a scene. A match move routine may be used to insert a synthetic object into one frame of a video sequence based on the pose and geometric information of the frame, and calculate all other required object views of the synthetic object for the remaining frames using the pose and geometric information acquired as a result of the multi-view three dimensional estimation routine.

Abstract: The present invention selectively applies one of VLC tables stored in a memory for encoding a coded block pattern of a macroblock according to the number of blocks having an object within the macroblock, the number of blocks obtained using shape information, thereby reducing the amount of data transmitted and increasing coding efficiency. The present invention also selectively applies one of VLD tables stored in a memory for decoding a coded block pattern of a macroblock according to the number of blocks having an object within the macroblock, the number of blocks obtained using shape information.

Abstract: A pixel interpolating device using the IM-GPDCT method and capable of preventing noise generation divides an original image into a plurality of blocks and picks up a block to be processed as a target block. Further, an image peripheral to the target block is extracted as a peripheral image. By using the peripheral image as an extension region, DCT transform is carried out. Thereafter, the peripheral image is magnified, and the IM-GPDCT processing is carried by using the magnified image.

Abstract: In a binary joint transform correlator, a Fourier plane binarization threshold is provided which alternates between even and odd rows and reduces noise in the output plane and improves discrimination between matched and unmatched prints. Periodically correlating with a known standard input and normalizing the output peak intensity prevents correlator performance drift from degrading fingerprint recognition performance. Experimental system performance is assessed statistically, in terms of the false pass rate seen when operating with a constant false alarm rate. By applying both techniques, a false pass improvement from 6.95% to 0.076% is experimentally demonstrated at a false alarm rate of 0.1 percent.

Abstract: A method for merging images is disclosed. The present invention may be implemented as a method for merging two or more images to produce a merged image using a mask. Each of these images has mutually corresponding pixels. In other words, each pixel value in an image has a spatially corresponding pixel in the other images. The mask has mask values, each mask value corresponding to a pixel in an image. The mask value is used to determine the corresponding pixel value of the merged image. One embodiment uses this mask value to set a corresponding pixel value of the merged image to a pixel value of one of the two or more images. Another embodiment uses the mask value as a weight. This weight is used to calculate the pixel value as a combination of corresponding pixel values of the two or more images. By using such a mask, a merged image may be formed avoiding artifacts resulting from decompressing a highly compressed image.

Abstract: A verification screen is displayed using information obtained by a prescan, fine scan conditions are computed on the basis of the print magnification and whether or not the density of an image obtained by the prescan is higher than a predetermined value is determined, so as to determine whether or not it is necessary to move to high density reading mode. If it is not necessary to move to the high density reading mode, a fine scan is carried out according to the fine scan condition. If it is necessary to move to the high density reading mode, an image is read at a slow transporting speed and over a long accumulation time.