Showing papers on "Image gradient published in 2000"

Image registration by maximization of combined mutual information and gradient information

Abstract: Mutual information has developed into an accurate measure for rigid and affine monomodality and multimodality image registration. The robustness of the measure is questionable, however. A possible reason for this is the absence of spatial information in the measure. The present paper proposes to include spatial information by combining mutual information with a term based on the image gradient of the images to be registered. The gradient term not only seeks to align locations of high gradient magnitude, but also aims for a similar orientation of the gradients at these locations. Results of combining both standard mutual information as well as a normalized measure are presented for rigid registration of three-dimensional clinical images [magnetic resonance (MR), computed tomography (CT), and positron emission tomography (PET)]. The results indicate that the combined measures yield a better registration function does mutual information or normalized mutual information per se. The registration functions are less sensitive to low sampling resolution, do not contain incorrect global maxima that are sometimes found in the mutual information function, and interpolation-induced local minima can be reduced. These characteristics yield the promise of more robust registration measures. The accuracy of the combined measures is similar to that of mutual information-based methods.

EdgeFlow: a technique for boundary detection and image segmentation

Abstract: A novel boundary detection scheme based on "edge flow" is proposed in this paper. This scheme utilizes a predictive coding model to identify the direction of change in color and texture at each image location at a given scale, and constructs an edge flow vector. By propagating the edge flow vectors, the boundaries can be detected at image locations which encounter two opposite directions of flow in the stable state. A user defined image scale is the only significant control parameter that is needed by the algorithm. The scheme facilitates integration of color and texture into a single framework for boundary detection. Segmentation results on a large and diverse collections of natural images are provided, demonstrating the usefulness of this method to content based image retrieval.

In search of illumination invariants

Abstract: We consider the problem of determining functions of an image of an object that are insensitive to illumination changes. We first show that for an object with Lambertian reflectance there are no discriminative functions that are invariant to illumination. This result leads as to adopt a probabilistic approach in which we analytically determine a probability distribution for the image gradient as a function of the surface's geometry and reflectance. Our distribution reveals that the direction of the image gradient is insensitive to changes in illumination direction. We verify this empirically by constructing a distribution for the image gradient from more than 20 million samples of gradients in a database of 1,280 images of 20 inanimate objects taken under varying lighting condition. Using this distribution we develop an illumination insensitive measure of image comparison and test it on the problem of face recognition.

Image searching method and image processing method

Abstract: The image searching method extracts or recognizes specific information for an image that exists in the image and/or accompanies it, allows either the specific information itself or related information thereof to be appendant to image data for the image, stores the image data as well as either information appendant to the image data as accessory information in a database, designates at least a portion of the accessory information as a searching condition, searches through the database under the designated searching condition, determines a degree of agreement with the accessory information appendant to a selected image data and reads an image having the degree of agreement not lower than a predetermined value. The image processing method utilizes the image searching method. These methods are capable of efficient image detection or search using as a key specific information such as the shape of subjects.

Object tracking using deformable templates

Abstract: We propose a method for object tracking using prototype-based deformable template models. To track an object in an image sequence, we use a criterion which combines two terms: the frame-to-frame deviations of the object shape and the fidelity of the modeled shape to the input image. The deformable template model utilizes the prior shape information which is extracted from the previous frames along with a systematic shape deformation scheme to model the object shape in a new frame. The following image information is used in the tracking process: 1) edge and gradient information: the object boundary consists of pixels with large image gradient, 2) region consistency: the same object region possesses consistent color and texture throughout the sequence, and 3) interframe motion: the boundary of a moving object is characterized by large interframe motion. The tracking proceeds by optimizing an objective function which combines both the shape deformation and the fidelity of the modeled shape to the current image (in terms of gradient, texture, and interframe motion). The inherent structure in the deformable template, together with region, motion, and image gradient cues, makes the proposed algorithm relatively insensitive to the adverse effects of weak image features and moderate amounts of occlusion.

The Manhattan World Assumption: Regularities in Scene Statistics which Enable Bayesian Inference

Abstract: Preliminary work by the authors made use of the so-called "Manhattan world" assumption about the scene statistics of city and indoor scenes. This assumption stated that such scenes were built on a cartesian grid which led to regularities in the image edge gradient statistics. In this paper we explore the general applicability of this assumption and show that, surprisingly, it holds in a large variety of less structured environments including rural scenes. This enables us, from a single image, to determine the orientation of the viewer relative to the scene structure and also to detect target objects which are not aligned with the grid. These inferences are performed using a Bayesian model with probability distributions (e.g. on the image gradient statistics) learnt from real data.

Consistent gradient operators

Abstract: We propose optimal gradient operators based on a newly derived consistency criterion. This criterion is based on an orthogonal decomposition of the difference between a continuous gradient and discrete gradients into the intrinsic smoothing effect and the self-inconsistency involved in the operator. We show that consistency assures the exactness of gradient direction of a locally 1D pattern in spite of its orientation, spectral composition, and sub-pixel translation. Stressing that inconsistency reduction is of primary importance, we derive an iterative algorithm which leads to accurate gradient operators of arbitrary size. We compute the optimum 3/spl times/3, 4/spl times/4, and 5/spl times/5 operators, compare them with conventional operators and examine the performance for one synthetic and several real images. The results indicate that the proposed operators are superior with respect to accuracy, bandwidth and isotropy.

Extending dynamic range of monochrome and color images through fusion

Abstract: This paper presents an approach to enhance the dynamic range of monochromatic and color images. The goal of this method is that of overcoming the inherent physical limitations of the acquisition process. The approach uses pattern-selective image fusion as a means to integrate information from images acquired under different shutter speeds or apertures. The result of this process is a composite in which none of the portions of the image is over or under saturated and the underlying information has been preserved. The results show the performance of this enhancement process from imagery taken from consumer-grade video camera equipment.

Image recognition method and apparatus utilizing edge detection based on magnitudes of color vectors expressing color attributes of respective pixels of color image

Abstract: An image recognition apparatus operates on data of a color image to obtain an edge image expressing the shapes of objects appearing in the color image, the apparatus including a section for expressing the color attributes of each pixel of the image as a color vector, in the form of a set of coordinates of an orthogonal color space, a section for applying predetermined arrays of numeric values as edge templates to derive for each pixel a number of edge vectors each corresponding to a specific edge direction, with each edge vector obtained as the difference between weighted vector sums of respective sets of color vectors of two sets of pixels which are disposed symmetrically opposing with respect to the corresponding edge direction, and a section for obtaining the maximum modulus of these edge vectors as a value of edge strength for the pixel which is being processed. By comparing the edge strength of a pixel with those of immediately adjacent pixels and with a predetermined threshold value, a decision can be reliably made for each pixel as to whether it is actually located on an edge and, if so, the direction of that edge.

Method of and system for image processing and recording medium for carrying out the method

Abstract: A processed image signal is obtained from an original image signal representing an original image by carrying out on the original image signal an image processing based on a signal representing information on a high frequency component of the original image signal. Band-limited image signals are made from the original image signal, and a signal representing information on a high frequency component of the original image signal is obtained from the band-limited image signals on the basis of a predetermined transformation function. The image processing is carried out on the original image signal on the basis of the signal representing information on a high frequency component of the original image signal.

Efficient image gradient based vehicle localization

Abstract: This paper reports novel algorithms for the efficient localization and recognition of traffic in traffic scenes. The algorithms eliminate the need for explicit symbolic feature extraction and matching. The pose and class of an object is determined by a form of voting and one-dimensional (1-D) correlations based directly on image gradient data, which can be computed "on the fly." The algorithms are therefore very well suited to real-time implementation. The algorithms make use of two a priori sources of knowledge about the scene and the objects expected: (1) the ground-plane constraint and (2) the fact that the overall shape of road vehicles is strongly rectilinear. Additional efficiency is derived from making the weak perspective assumption. These assumptions are valid in the road traffic application domain. The algorithms are demonstrated and tested using routine outdoor traffic images. Success with a variety of vehicles in several traffic scenes demonstrates the efficiency and robustness of context-based image understanding in road traffic scene analysis. The limitations of the algorithms are also addressed.

Image processing apparatus, image processing method, and recording medium with image processing program to process image according to input image

Abstract: An image processing apparatus includes a reception unit receiving image data, a region determination unit to extract a text region from the received image data, a threshold value calculation unit to calculate the background amount of the extracted text region, and an image correction unit to correct a non-text region of the image data according to the calculated background amount. Since the non-text region of the image data is corrected according to the background amount of the text region, an appropriate correction process can be applied on the image data according to difference in the paper sheet quality and image pickup condition.

Converting a multi-pixel image to a reduced-pixel image to provide an output image with improved image quality

Abstract: A CCD image sensor 14 uses sufficiently more pixels than those in an output image. A pixel number conversion block applies pixel number reduction conversion to an image pickup signal from the CCD image sensor 14. This signal is output as a monitor output or motion-picture recording signal via a terminal 43. Converting the multi-pixel image signal to a reduced image can provide an output image with improved sharpness.

Digital image processing method for edge shaping

Abstract: A digital image processing method, includes the steps of: transforming the digital image using an edge sensitive wavelet transform to produce a plurality of wavelet coefficients at various resolutions and a residual image; modifying the wavelet coefficients as a function of the image gradient and its rate of change at a resolution corresponding to the respective wavelet coefficient; and inverse transforming the modified wavelet coefficients and the residual image to produce a processed digital image.

Image pickup system, image processor, and camera

Abstract: An image capturing system, image processing system, and camera in which color can be precisely corrected to realize invariability of color through simple calibration while decreasing the area of a reference image part. The system comprises a camera having a lens, an image capturing system device, and a reflecting surface. Reference signal values (rn, gn, bn) are determined by averaging the intensities of the reflected lights of a reference scene received by the image capturing system device from the respective color channels of a plurality of pixel parts. The reference signal values represent the light source color, and a main image is corrected using the reference signal values.

System and method for performing region-based image retrieval using color-based segmentation.

Abstract: There is disclosed an image retrieval system for analyzing an image in a first color model format and detecting and retrieving from the image a selected image portion. The image retrieval system comprises an image processor for converting pixels in the image from the first color model format to a (Yrυ) color model format, where Y is an intensity component indicating a total amount of light, r is a saturation component indicating an amount of white light mixed with a color of each pixel, and υ is a hue component indicating the color of each pixel. The image processor groups spatially adjacent pixels into image regions according to hue components of the adjacent pixels and performs a merging process wherein a first image region and an adjacent second image region are merged into a composite region if a hue difference between the first and second image regions is less than a predetermined hue difference threshold. The process is repeated to continually merge regions of similar hue until no further merges can be performed.

Tomographic image reading method, automatic alignment method, apparatus and computer readable medium

Abstract: A tomographic image reading method for extracting a comparison image corresponding to a diagnostic image, the diagnostic image being one of first tomographic images, the comparison image being one of second tomographic images, the method including the steps of: inputting the first images and the second images; generating a first projection image from the first images and a second projection image from the second images; measuring shift amount between the first projection image and the second projection image by using a template; correcting the slice position according to the shift amount; and displaying the diagnostic image and the comparison image to a monitor.

Method, apparatus and recording medium for image processing

Abstract: The color and tone of an image of an object are changed to a color and tone desired by a user. Image data including an image of an object is input to image processing means and displayed on a monitor. The user observes the image displayed on the monitor and selects an area including an image of an object having the desired color and tone, and an area including the image of an object whose color and tone are to be changed to the desired color and tone. The image processing means extracts color pixels from each respective specified area and generates cumulative histograms of the color pixels as characteristic quantitative data representing the color and tone. Based on the cumulative histograms, the color-tone of the area including the image of the latter object is changed, and processed image data whose color and tone have been changed is obtained.

Anisotropic edge-preserving smoothing in carotid B-mode ultrasound for improved segmentation and intima-media thickness (IMT) measurement

Abstract: Non-invasive B-mode ultrasound imaging of carotid artery intima-media thickness (IMT) is used for estimation of the atherosclerotic burden. Image interpretation and measurement is degraded by imaging noise and other artifacts. Edge detection based on local estimation of image gradient produces irregular boundaries and poor IMT estimates under such conditions, Manual tracing produces smoother borders but is labor-intensive and exhibits poorer intra and inter-observer reproducibility. We present a new non-linear segmentation algorithm that uses global image information to improve border delineation, simplify image frame selection and improve IMT measurement We demonstrate examples of image enhancement and evaluate the measurement reproducibility of this method in 40 carotid data sets from an ongoing clinical study.

Image supervisory method and device

Abstract: PROBLEM TO BE SOLVED: To provide an image supervisory device that monitors a dark scene with an extended dynamic range and a reduced lighting power. SOLUTION: An image input control section 500 sets a 1st image with normal exposure at strong lighting and a 2nd image with high-speed exposure at weakend lighting to a supervisory scene photographed by an ITV camera 100 with an LED lighting device 200, and an image input section 1000 receives the 1st image and then receives a 2nd image at a succeeding field. An image synthesis section 2000 sums the 1st and 2nd images with a weight coefficient corresponding to the luminance by each pixel multiplied therewith to generate a composite image thereby obtaining a processing object image for object discrimination. A differential image generating section 3000 conducts differential processing by each pixel between a just preceding frame and a current frame of the composite image, and a change area extract section 4000 applies binary processing to the differential image to extract a change area. An object detection section 5000 conducts contrast pattern matching for each inter-frame normalization processing per integrated change area to discriminate and object.

Pose estimation, model refinement, and enhanced visualization using video

Abstract: In this paper we present methods for exploitation and enhanced visualization of video given a prior coarse untextured polyhedral model of a scene. Since it is necessary to estimate the 3D poses of the moving camera, we develop an algorithm where tracked features are used to predict the pose between frames and the predicted poses are refined by a coarse to fine process of aligning projected 3D model line segments to oriented image gradient energy pyramids. The estimated poses can be used to update the model with information derived from video, and to re-project and visualize the video from different points of view with a larger scene context. Via image registration, we update the placement of objects in the model and the 3D shape of new or erroneously modeled objects, then map video texture to the model. Experimental results are presented for long aerial and ground level videos of a large-scale urban scene.

3-D image analysis of abdominal aortic aneurysm.

Abstract: In this paper we propose a technique for 3-D segmentation of abdominal aortic aneurysm (AAA) from computed tomography (CT) angiography images. Output data form the proposed method can be used for measurement of aortic shape and dimensions. Knowledge of aortic shape and size is very important for selection of appropriate stent graft device for treatment of AAA. The technique is based on a 3-D deformable model and utilizes the level-set algorithm for implementation of the method. The method performs 3-D segmentation of CT images and extracts a 3-D model of aortic wall. Once the 3-D model of aortic wall is available it is easy to perform all required measurements for appropriate stent graft selection. The method proposed in this paper uses the level-set algorithm instead of the classical active contour algorithm developed by Kass et al. The main advantage of the level set algorithm is that it enables easy segmentation surpassing most of the drawbacks of the classical approach. In the level-set approach for shape modeling, a 3-D surface is represented by a real 3-D function or equivalent 4-D surface. The 4-D surface is then evolved through an iterative process of solving the differential equation of surface motion. Surface motion is defined by velocity at each point. The velocity is a sum of constant velocity and curvature-dependent velocity. The stopping criterion is calculated based on image gradient. The algorithm has been implemented in MATLAB and C languages. Experiments have been performed using real patient CT angiography images and have shown good results.

Image processing device and image processing method for correction of image distortion

Abstract: For finding an inclination of an image rapidly and correctly, an image processing device includes a CCD picking up an image, an edge detection unit detecting an edge of an image relating to the picked-up image, a rotation unit rotating the detected edge in an image plane, an operation unit deriving a characteristic amount of rotated edge and an inclination detection unit detecting an inclination of the picked-up image based on the derived characteristic amount.

Balanced template tracker for tracking an object image sequence

Abstract: A method and apparatus are described for tracking an object image in an image sequence in which a template window associated with the object image is established from a first image in the image sequence and an edge gradient direction extracted. A search window associated with the object image is established from a second image in the image sequence and a second edge gradient direction associated with the search window extracted. The first and second edge gradient directions are weighted and correlated to provide a coordinate offset between relative positions of the object image in the search and template windows to track the object image in the image sequence. If the coordinate offset exceeds a predetermined value the template window is updated by establishing a new template window. Weighting further includes calculating a set of weighting coefficients for the edge gradient direction associated with the search window and the template window respectively such that a balanced contribution is provided from respective edge gradient directions in correlation. The edge direction gradient associated with the template window is auto-correlated to determine a quality measure. Extracting the second edge gradient direction further includes I filtering the second edge gradient direction and an edge magnitude, thinning the second edge gradient direction, thresholding the edge magnitude, and normalizing the second edge gradient direction if the edge magnitude is successfully thresholded.

Wavelet-Based Salient Points: Applications to Image Retrieval Using Color and Texture Features

Abstract: In image retrieval, global features related to color or texture are commonly used to describe the image. The use of interest points in content-based image retrieval allows image index to represent local properties of images. Classic corner detectors can be used for this purpose. However, they have drawbacks when applied to various natural images for image retrieval, because visual features need not to be corners and corners may gather in small regions. We present a salient point detector that extracts points where variations occur in the image, regardless whether they are corner-like or not. It is based on wavelet transform to detect global variations as well as local ones. We show that extracting the color information in the locations given by these points provides significantly improved retrieval results as compared to the global color feature approach. We also show an image retrieval experiment based on texture features where our detector provides better retrieval performance comparing with other point detectors.

Image processing device, image processing method, and recorded medium

Abstract: An image processing device in which the storage size for image data inputted from a scanner or an electronic still camera or transmitted through communication means such as the Internet is small, and a clear magnified image can be outputted. To reduce the storage size for the image data, low-frequency components extracting means extracts the low-frequency component in the image data, and encodes the high-frequency component after finding information concerning the low-frequency component. To output a clear magnified image, magnified image frequency estimating means estimates the frequency components necessary for image-magnifying by edge extraction and by nonlinear interpolation. Image magnifying means estimates a magnified image by wavelet conversion.

Recovering the 3D shape of a road by on-board monocular vision

Abstract: Deals with a method designed to recover the 3D geometry of a road from an image sequence provided by an on-board monocular monochromatic camera. It only requires the road edges to be detected in the image. The reconstruction process is able to compute (1) the 3D coordinates of the road axis points, (2) the vehicle's position on its lane and (3) the prediction of the road edge localization in the next images of the sequence which is very helpful for the detection phase. It also computes the confidence intervals associated with the 3D parameters. The description of the method is followed by the presentation of its most significant results.

Two practical issues in Canny's edge detector implementation

Abstract: We address two practical issues, namely smoothing factor selection and efficient implementation of thresholding with hysteresis, in implementing Canny's edge detector. The smoothing factor of the Gaussian kernel should be chosen to maximize the discrete version of Canny's original criteria. Thresholding with hysteresis should be implemented using an efficient connected component analysis algorithm. Following these suggestions in implementing Canny's edge detector will in general result in optimal edge detection quality and very significant reduction in running time for large images.