A method for automated detection of abnormal anatomic regions, wherein a mammogram is digitized to produce a digital image and the digital image is processed using local edge gradient analysis and linear pattern analysis in addition to feature extraction routines to identify abnormal anatomic regions. Noise reduction filtering and pit-filling/spike-removal filtering techniques are also provided. Multiple difference imaging techniques are also used in which difference images employing different filter characteristics are obtained and processing results logically OR'ed to identify abnormal anatomic regions. In another embodiment the processing results with and without noise reduction filtering are logically AND'ed to improve detection sensitivity. Also, in another embodiment the wavelet transform is utilized in the identification and detection of abnormal regions. The wavelet transform is preferably used in conjunction with the difference imaging technique with the results of the two techniques being logically OR'ed.

A method of computerized analysis of temporally sequential digital images, including (a) determining first shift values between pixels of a first digital image and corresponding pixels of a second digital image; (b) warping the second digital image based on the first shift values to obtain a first warped image in which spatial locations of pixels are varied in relation to the first shift values; (c) determining second shift values between pixels of the first digital image and pixels of the first warped image; and (d) warping the first warped image based on the second shift values to obtain a second warped image in which spatial locations of pixels of the first warped image are varied in relation to the second shift values. Additional iterations of image warping are possible to enhance image registration between the first digital image and the warped version of the second digital image, followed by image subtraction of the first digital image and the final warped image to produce a difference image from which diagnosis of temporal changes ensues. Temporal subtraction assists radiologists in the detection of interval changes on chest radiographs, and particularly to overcome severe misregistration errors in the temporally sequential images mainly due to differences in a subject's inclination and/or rotation. In the production of shift values used in image warping, initial shift values are obtained by cross-correlation techniques using a template and search regions of interest. Shift vectors and a histogram of shift vectors in each lung are obtained from initial shift values. The histograms of shift vectors are used in the selection of sets of shift values for smoothing, using two-dimensional fitting and subsequent use of fitted shift values in image warping.

A morphology operation is carried out on an original image signal representing a radiation image, and a morphology signal is thereby obtained. An edge signal, which represents a characteristic value with respect to an image edge portion in the radiation image, or a specific image signal, which represents a characteristic value with respect to a specific image portion having a contour of a predetermined size, is obtained from the morphology signal and the original image signal. An emphasis coefficient corresponding to the edge signal or the specific image signal is calculated from a conversion table. The conversion table is set such that the emphasis coefficient may take a positive value when the value of the edge signal or the specific image signal is larger than a predetermined threshold value. The conversion table is corrected such that, as a dose of radiation delivered during an operation for recording the radiation image becomes small, the threshold value may become large.

Abnormal pattern detection processing is performed on an entire area image signal, which represents an entire area image, and an abnormal pattern embedded in the entire area image is detected. A local area limited region, which is constituted of the detected abnormal pattern and a region neighboring with the abnormal pattern, is set as a region-of-interest image. Information defined in one of (1), (2), (3), and (4) below is fed into an external image display terminal: (1) the entire area image signal and a detection result signal, which represents detection results of the abnormal pattern having been detected, (2) the entire area image signal, the detection result signal, and a region-of-interest image signal, which represents the region-of-interest image, (3) an embedding entire area image signal representing an embedding entire area image, which is formed by embedding the detection results of the abnormal pattern into the entire area image, and (4) the embedding entire area image signal and the region-of-interest image signal.

An image classification apparatus includes an image input device for inputting image data. A filing device stores the input image data and performs a read, retrieval, or edit operation on the data in accordance with a predetermined instruction. A normalizing unit corrects variations in various image input conditions or variations in image input devices to make the input conditions in execution agree with those in learning. A feature extracting unit extracts a feature amount effective for classification. A classification determination unit performs a classification determination operation on the basis of the feature amount obtained by the feature extracting unit. A display unit synthesizes a result of the classification determination operation with the input image, and displays the synthesized result. A learning control unit controls the feature extracting unit and the classification determination unit on the basis of the predetermined learning data. Accordingly, the classification apparatus classifies a portion of the input image which is difficult to extract by binarization or three-dimensional display alone, and displays the classified portion in an easily recognizable visual form.