Eastman-kodak Company

Carestream Health, Inc.

Lori L Barski

Roger S Gaborski

Robert Allen Senn

Xiaohui Wang

Mary E Couwenhoven

Hin-Leong Tan

William J Sehnert

Hui Luo

Levon O Vogelsang

Hsien-Che Lee

Louis J Beato

Karin Toepfer

Lynn M Fletcher-Heath

Andrew M Assad

Richard L Van Metter

Dawn L Dutton

A machine learning network is trained to generate a noise field image from a radiographic (x-ray) image. The training includes accessing a number of previously acquired radiographic images, duplicating the previously acquired radiographic images, and conditioning each of the duplicated images with simulated noise content to form a plurality of simulated low-exposure images. Each of the simulated low-exposure images is paired with its corresponding previously acquired image to form a learning pair. The machine learning network is trained to generate a noise field image using the learning pairs of images. A noise suppressed image of an object can be generated by applying a scaling factor to at least a portion of the corresponding noise field image and combining the scaled noise field image with a current captured image of the object.

Noise suppression using deep convolutional networks

A method for rendering volume radiographic image content of a subject forms a volume image. The method extracts a first image slice from the volume image, then modifies the extracted first image slice by defining two or more spatial frequency bands from the image slice data and applying one or more viewer adjustments to the image slice data, wherein the one or more viewer adjustments condition the image data to enhance image content in at least one of the defined spatial frequency bands. A set of display rendering parameters is generated according to the two or more frequency bands and according to viewer adjustments made for the first image slice. A second image slice is extracted from the volume image. The generated set of display rendering parameters is applied to the second image slice to render an adjusted image slice and the adjusted image slice is displayed.

Method for enhanced display of image slices from 3-D volume image

A method for classifying tissue density of a breast includes obtaining mammography image data and segmenting the mammography image to identify the region representing the breast tissue. A plurality of regions within the breast tissue region are identified for obtaining image features therefrom. A plurality of image features are computed from the identified plurality of regions. The breast tissue density is classified using the computed plurality of image features.

Method for classifying breast tissue density using computed image features

A method of displaying a digital medical image. The steps include acquiring a digital medical image; determining from the acquired digital medical image the anatomically relevant region which defines the relevant image boundaries; and determining an optimum virtual plate size from a stored plurality of virtual plate sizes for displaying the anatomically relevant regions of the digital medical image on a display device. In one arrangement, the acquired digital medical image has associated exam information (such as bodypart and image projection), the plurality of virtual plate sizes are stored by exam information, and the optimum plate size is determined, at least in part, by the exam information associated with the digital medical image.

Method for producing a cropped medical image

Method of automatically determining tone-scale parameters for a digital image

We demonstrate a comprehensive image processing system to automatically detect the use of a stationary anti-scatter device (also referred to as a linear grid) and suppress the artifacts that such a device creates in a digital radiographic images. The detection process consists of the stages of determining an appropriate region of the image from which to perform the analysis; performing a spectral analysis to identify the grid frequencies; and identifying the most likely grid line frequency. When a grid is detected, the results of the detection process are used to perform adaptive grid suppression using custom designed blurring filters and a plurality of adaptation methods.

Method for x-ray antiscatter grid detection and suppression in digital radiography

A method of determining the skin-line in a digital medical image comprising the steps of:

Detection of skin-line transition in digital medical imaging

A method of automatically adjusting the tone scale adjustment for digital radiographic images. The method includes the steps of providing an input digital radiographic image; estimating the image noise as a function of code value of the image; determining the range of code values in the input image; determining which way the input image should be processed; determining a bounding box which contains the region of interest of the input image; computing the image activity histogram of the input image; determining four points from the image activity histogram; constructing the tone-scale curve for the input image; and mapping the input image through the tone-scale curve to produce an output image with good tone scale.

Automatic tone scale adjustment using image activity measures

A method of determining direct x-ray exposure regions in digital medical imaging comprising the steps of: providing a digital medical image including a matrix of lines and columns of pixels; evaluating significant transitions in line and column profiles to determine appropriate line and column thresholds; evaluating all pixels which exceed the line/column threshold for structure by computing the variation within a moving window using a range statistic; accumulating within a histogram pixels which exceed a threshold for that particular line or column and which contain variation indicative of direct x-ray exposure; and analyzing the background histogram to determine the beginning intensity or code value (left point) of the background region.

Determination of direct x-ray exposure regions in digital medical imaging

The text to be classified as a dot matrix or ink jet printer character is scanned at a suitable resolution. Several horizontal and vertical slices are made through the bitmap image of individual characters. Density values of pixels contained in these slices are stored as a density profile for each slice. As a result of these density profiles, a determination is made as to whether or not the character was printed on a dot matrix or ink jet printer, or whether the characters were printed using a higher quality device such as a daisy wheel printer.

Method for detecting ink jet or dot matrix printing

A method and apparatus for performing cursive script recognition is disclosed, wherein a cursive word, in the form of digitized data or bitmap, is simultaneously segmented and individual characters of the word are recognized using a scanning window that moves across a word field or segment. The bitmap data is preprocessed before being presented to a moving window letter center finding neural network that determines the spatial location of the centers of individual letters. Character center data generated by the center finding neural network is then used to define the left and right edges of a fixed size window. The fixed window contains width normalized word segments that are presented to a second neural network which is taught to recognize the central character of the window based on portions of adjacent characters contained within the window. A final character recognition decision is made using a third neural network that combines the results of second neural network at three positions, namely, when the character to be recognized was in a precharacter position, a central position and a post character position relative to the third neural network.

Method and apparatus for cursive script recognition

An image processing technique especially useful in processing digital radiographic images. A method for finding a histogram region of interest for improved tone scale reproduction of digital radiographic images includes the following steps. A digital radiographic image is randomly sampled with a sample having an appropriate size to delineate an object of interest. Each sample is processed using texture analysis techniques to extract a plurality of texture features. Using the extracted texture features, each sample is classified with a previously trained neural network classifier to determine its class. Last, the pixel values belonging to the same class are accumulated to form separate histograms for each class. Each of the histograms are then used to optimize tone scale reproduction.

Technique for finding the histogram region of interest for improved tone

Character segmentation apparatus for an optical character recognition system for segmenting individual character images in an image of a document having many characters prior to performing character identification, including a movable kernel for capturing a sub-image framed within a window having an area corresponding to an area occupied by an individual character, the window being movable in the document image in pixel-by-pixel steps to capture a sub-image for each step of the window, an associative memory for responding to the captured sub-image by producing a corresponding one of a set of images of known characters with which the associative memory has been trained, and a sensor responsive to the behavior of the associative memory for determining whether the sub-image is the image of an individual character or a non-character.

Character segmentation using an associative memory for optical character

Method and apparatus are disclosed for processing image data of dot-matrix/ink-jet printed text to perform Optical Character Recognition (OCR) of such image data. In the method and apparatus, the image data is viewed for detecting if dot-matrix/ink-jet printed text is present. Any detected dot-matrix/ink-jet produced text is then pre-processed by determining the image characteristic thereof by forming a histogram of pixel density values in the image data. A 2-D spatial averaging operation as a second pre-processing step smooths the dots of the characters into strokes and reduces the dynamic range of the image data. The resultant spatially averaged image data is then contrast stretched in a third pre-processing step to darken dark regions of the image data and lighten light regions of the image data. Edge enhancement is then applied to the contrast stretched image data in a fourth pre-processing step to bring out higher frequency line details. The edge enhanced image data is then binarized and applied to a dot-matrix/ink jet neural network classifier for recognizing characters in the binarized image data from a predetermined set of symbols prior to OCR.

Growing community of inventors

Mendon, NY, United States of America

Lori L Barski