Canon Kabushiki Kaisha

Canon Inc.

Canon Research Center America, Inc.

Harry T Garland

Gerald A May

Roger D Melen

Michael G Roberts

Hadar I Avi-Itzhak

Alexander Berestov

Kenneth M Hunter

Thanh A Diep

Olagappan Manickam

Ken Hunter

Jonathan DeVito

David D Deer

A radiographic imaging system ( ) comprises an X-ray tube ( ), a sensor plate ( ), and a graphics engine ( ). The tube ( ) and the sensor plate ( ) rotate synchronously about a patient ( ) and expose a stereoscopic pair of images which are transmitted to the graphics engine ( ). The graphics engine ( ) determines ( ) the geometry of the system ( ). If ( ) the pair of images are toed-in relative to each other, the graphics engine ( ) converts ( ) the images into a parallel geometry. Likewise, the graphics engine ( ) also processes ( ) the images for keystone distortion, if necessary. Simply flipping the images in the stereo pair distorts the depth of objects in the stereoscopic image. Instead of simply flipping the images, it is desirable to &ldquo;go behind&rdquo; the screen ( A) and look at the image from the back. In order to produce a stereoscopic image having this viewpoint, the graphics engine ( ) uses the geometry of the radiographic system ( ) to adjust ( ) the screen parallax of the image after the constituent images are flipped ( ) and provide the desired three-dimensional perspective.

Flipping stereoscopic radiographs

A computer-implemented method and system for adjustment and matching of stereoscopic images ( ) of the eye fundus ( ). Illumination errors between the images are eliminated, such as by adjusting groups of pixels so that the mean and variance of a group ( ) from the left image ( ) will have the same mean intensity and variance as the corresponding group ( ) from the right image ( ). An epipolar line adjustment is performed, such as by shifting points in the images based on matched points in the images. Image occlusion errors are detected and removed, such as by matching point analysis. Once these adjustment steps have been completed, a matching analysis can be preformed using the corrected images.

Automated stereo fundus evaluation

A self-diagnosing image sensor ( ) detects and stores maps of functioning and malfunctioning pixels ( ) in a memory ( ) directly coupled to the sensor ( ). The memory ( ) is coupled to an external monitoring computer ( ) which retrieves the pixel map and adjusts the sensor data received from the image sensor ( ) in accordance with the retrieved pixel map. A defect discriminator ( ) is coupled directly to the image sensor ( ) and to the memory ( ) for detecting whether a pixel ( ) malfunctions, and updates the map accordingly. Additionally, if the number of malfunctioning pixels ( ) in the sensor ( ) exceeds a predefined threshold, an alert message is available to the external monitoring computer or display ( ) to warn the user that the sensor ( ) may be generating inaccurate information. An on-plate pixel processor ( ) performs any necessary interpolation of the date responsive to the pixel map, and a complete image is sent to the remote display ( ), without requiring any further processing.

Self-diagnostic image sensor

A system and method for automatically collimating X-rays. A digital X-ray system ( ) includes a generator ( ), a sensor unit ( ), a control station ( ), and a preview monitor ( ). The generator ( ) generates X-ray radiation that is captured by the sensor unit ( ) as a digital image and transmitted to the control station ( ). The captured image is displayed on the preview monitor ( ). The generator ( ) includes a collimator ( ) that collimates the generated radiation into a primary beam of X-rays. The size and shape of the primary beam can be adjusted by modifying collimation parameters. A short duration beam of X-rays is generated by the generator ( ) and captured ( ) by the sensor unit ( ). This step is repeated as necessary or desired. The resulting digital images are analyzed ( ) by the control station ( ) to calculate a calibration coefficient. Another short duration beam of X-rays is generated and a reference image is captured ( ). The control station ( ) analyzes the reference image and uses the calibration coefficient to collimate the collimator ( ) to achieve a desired goal, such as radiating all or only part of an object being radiated. Once the collimator ( ) is adjusted, the X-ray image of the subject is exposed.

Auto-collimating digital X-ray system

An integrated digital television and video printer ( ) comprises a first decompressor ( ) for decompressing a compressed digital video signal corresponding to a real time broadcast, a buffer ( ) for storing a compressed digital video signal ( ), a synchronizing means ( ) for coordinating the display of the real time digital video signal and the stored digital video signal, a second decompressor ( ) for decompressing the stored compressed digital video signal received from the buffer ( ), a control means ( ) for selecting among and printing images corresponding to the digital video signal stored in the buffer ( ), a display means ( ) for viewing the real time digital video signal and/or the stored digital video signal, and a printing means ( ) for producing a hard copy of the selected image from the buffer ( ).

Integrated digital television and video printer

Diagnostic x-ray images having uniform optical density are generated by selectively scanning (,) lines of pixels (,) in a sensor unit (,) used to capture x-ray images in a digital radiography system (,). By selectively scanning (,) the lines of pixels (,), the exposure time of such pixels can be controlled. The lines of pixels (,) are selectively scanned (,) according to exposure compensation profiles (,) for the anatomic regions of interest. The exposure compensation profiles (,) include an ordered sequence of exposure times which can be loaded into a programmable timer (,) coupled to read out circuitry (,) located in the sensor unit (,). The timer (,) uses this ordered sequence of exposure times to determine which pixel line to access next, and when to initiate that access. The exposure compensation profiles (,) can be generated (,) from empirical data by taking (,) preliminary low dose exposures of the body parts to be examined.

Exposure compensation for digital radiography systems using selective scanning of sensor arrays

A method, computer program product, and digital radiography system (,) detect type information concerning an anti-scatter grid (,), and store the information in a header (,) of a file (,) containing a digital x-ray image. A gleaning module (,) gleans relevant information from an identifier (,) on a physical medium (,) containing the anti-scatter grid (,). A retrieval module (,) retrieves an entry (,) corresponding to information gleaned from the identifier (,) and containing information concerning the anti-scatter grid (,) from a look-up table (,) in a computer memory (,). A storage module (,) stores the information in a header (,) of a file (,) containing the digital x-ray image. Additionally, a comparison module (,) compares type information concerning the identified anti-scatter grid (,) with type information for a correct anti-scatter grid for a procedure being performed. When the anti-scatter grid (,) being used is incorrect for the procedure, an alerting module (,) alerts an operator of the digital radiography system (,) of the error.

Automatic grid parameter logging for digital radiography

A method and system for enhancing a digital image stores the original image and a series of codes describing the manipulations performed on the image. Each time the image is retrieved from an original image archive (,) for viewing on a display (,), corresponding codes are retrieved from an original sequence archive (,). If the codes have previously been optimized, the image processing system (,) applies the optimized codes to the image in order to recreate the enhanced image. If the codes have not been optimized, an optimizer (,) combines and reorders the image enhancement codes in light of certain criteria, including minimizing information loss. A strategy used by the optimizer (,) to minimize information loss separates the codes into sequences and then commutes and combines the codes based on the code types. The optimized codes are applied to the image and produce the image that is displayed on the display (,) and saved in the enhanced image cache memory (,).

Optimized enhancement of digital images

A digital x-ray image (,) is divided (,) into regions (,), (,) according to a selected (,) exposure compensation profile. Each region (,),(,) includes pixel values indicative of gray-scale levels in that region. The pixel values are used to index entries in one of a plurality of spatial LUTs (,) assigned (,) to that region. Entries in the selected spatial LUT (,) contain new pixel values indicative of a transform (,) on the original pixel values used to index the spatial LUT (,). Original pixel values are replaced (,) with corresponding pixel values from the spatial LUT (,). Each region of the x-ray image (,) needing compensation is transformed (,), in turn, until the entire image (,) is processed.

Exposure compensation for digital radiography systems using spatial look-up tables

The x-ray exposure of an image pixel array (,) is measured by assigning to certain pixels (,) of the array the task of measuring the level of exposure. A target set (,) of pixels from array (,) is selected and divided into a set of exposure pixels (,) and a set of image pixels (,). Both sets of pixels generate and collect electrical charges in proportion to the amount of x-ray exposure experienced by them. The charges collected by the exposure pixels (,) are continuously read out during the capture of an image and compared with a desired exposure level. Once the desired exposure level is reached, as indicated by the exposure pixels (,), the image pixels (,) are read out and an image is produced.

Exposure control for digital radiography systems using charge build-up in sensor array pixels

Regions of a source image (900) are encoded at different levels of image quality. A plurality of image quality levels are defined (1010). In addition, the source image (900) is divided into one or more regions (912, 914, 916) and one of the image quality levels is assigned to each region (1012). Next, the entire source image (900) is encoded, preferably with JPEG encoding, at the lowest image quality level Q.sub.1, assigned to the image (1016). Then, the encoded source image (900) is decoded by a matching decoding process into a reference frame (1018). If the next highest image quality level Q.sub.2 is lossy, then the reference frame and source image are scaled according to the image quality level Q.sub.2 (1020). Then, differences between the source image and the reference frame are determined for all lossy regions assigned an image quality level greater than or equal to the image quality level Q.sub.2, and stored in a differential frame (1024). The differential frame is encoded according to the image quality level Q.sub.2 (1026), stored with the reference frame, then decoded to become the new reference frame (1028). These steps are repeated until the highest quality level assigned to a region is encoded. If the image quality level is lossless, then the reference frame is scaled to the resolution of the source image (1030), and the differential components are losslessy encoded (1038). When the variably encoded image is decoded, each of the selected regions (912, 914, 916) has the image quality defined by the assigned image quality level.

Variable compression encoding of digitized images

A personal imaging computer system, which is connectable to and operable with a computerized local or wide area network, identifies characters in a document on which the characters are formed. The system scans the document to obtain a gray-scale image of the document, generates a binary image from the gray-scale image by comparing the gray-scale image with the threshold, segments the binary image to locate individual characters within the binary image and to determine the shape of the individual characters, extracts gray-scale image information from the gray-scale image for each such individual character based on the location and shape of the character in the binary image, recognition-processes the extracted gray scale image information to determine the identity of the character, and stores the identity of the character. The character identities are stored in a character file in association with the image so as to facilitate image retrieval with a text-based search over many such character files. Preferably, the image is not stored in its scanned-in resolution, but rather is stored at a reduced resolution so as to reduce memory storage requirements.

Reduced resolution document storage and retrieval system

A method and system for storing and selectively retrieving information, such as words, from a document set. The method includes generating an image data set representative of the information contained in the document set. The method also involves generating a text data set representative of a text portion of the information contained in the document set. A text-image correspondence (TIC) table is generated that includes data representative of coordinates information corresponding to each phrase of the document set. A search phrase is identified in response to user-specified search criteria and the search phrase is identified in the text image data set. Then, the TIC table is used to identify the coordinates information corresponding to the search phrase identified in the text data set. A display of the portion of the page containing the search phrase is generated using the coordinates information.

Selective document retrieval method and system

Pattern recognition, for instance optical character recognition, is achieved by training a neural network, scanning an image, segmenting the image to detect a pattern, preprocessing the detected pattern, and applying the preprocessed detected pattern to the trained neural network. The preprocessing includes determining a centroid of the pattern and centrally positioning the centroid in a frame containing the pattern. The training of the neural network includes randomly displacing template patterns within frames before applying the template patterns to the neural network.

Training a neural network using centroid dithering by randomly

Method and system for selectively retrieving information contained in a stored document set using a metric-based or 'fuzzy' finite-state non-deterministic automaton. An automaton is constructed (501) corresponding to a text string query, text strings are read (502) from storage and corresponding dissimilarity values are generated (505). Those strings resulting in values less than a given threshold are recorded (508) and listed for the user. Dissimilarity values are determined based on penalties associated with missing characters, extra characters, incorrect characters, and other differences between the text string query and a text string read from storage.

Growing community of inventors

Los Altos Hills, CA, United States of America

Harry T Garland