Showing papers on "Orientation (computer vision) published in 2003"

Recognising panoramas

Abstract: The problem considered in this paper is the fully automatic construction of panoramas. Fundamentally, this problem requires recognition, as we need to know which parts of the panorama join up. Previous approaches have used human input or restrictions on the image sequence for the matching step. In this work we use object recognition techniques based on invariant local features to select matching images, and a probabilistic model for verification. Because of this our method is insensitive to the ordering, orientation, scale and illumination of the images. It is also insensitive to 'noise' images which are not part of the panorama at all, that is, it recognises panoramas. This suggests a useful application for photographers: the system takes as input the images on an entire flash card or film, recognises images that form part of a panorama, and stitches them with no user input whatsoever.

A framework for a streamline-based probabilistic index of connectivity (PICo) using a structural interpretation of MRI diffusion measurements.

Abstract: PURPOSE: To establish a general methodology for quantifying streamline-based diffusion fiber tracking methods in terms of probability of connection between points and/or regions. MATERIALS AND METHODS: The commonly used streamline approach is adapted to exploit the uncertainty in the orientation of the principal direction of diffusion defined for each image voxel. Running the streamline process repeatedly using Monte Carlo methods to exploit this inherent uncertainty generates maps of connection probability. Uncertainty is defined by interpreting the shape of the diffusion orientation profile provided by the diffusion tensor in terms of the underlying microstructure. RESULTS: Two candidates for describing the uncertainty in the diffusion tensor are proposed and maps of probability of connection to chosen start points or regions are generated in a number of major tracts. CONCLUSION: The methods presented provide a generic framework for utilizing streamline methods to generate probabilistic maps of connectivity.

Block Adjustment of High-Resolution Satellite Images Described by Rational Polynomials

Abstract: model provides a rigorous, accurate method to block adjust This paper describes how to block adjust high-resolution Ikonos data outside of the ground stations. satellite imagery described by Rational Polynomial Coefficient This publication of a technique for block adjusting Ikonos (RPC) camera models and illustrates the method with an Ikonos images described by RPC data is motivated by a desire to satisfy example. By incorporating a priori constraints into the the needs of those users who would like to perform their own adjustment model, multiple independent images can be block adjustment, and to ensure that Ikonos images are proadjusted with or without ground control. The RPC block cessed in such way as to consistently achieve the highest possiadjustment model presented in this paper is directly related ble accuracy. In developing the adjustment model described to geometric properties of the physical camera model. Multiple here, the authors had access to the complete description of physical camera model parameters having the same net effect Ikonos imaging geometry, familiarity with all of the satellite on the object-image relationship are replaced by a single maneuvering modes, the resources of extensive test ranges and adjustment parameter. Consequently, the proposed method is imagery with which to test and validate, and the experience numerically more stable than the traditional adjustment of gained calibrating, testing, and troubleshooting Ikonos metric exterior and interior orientation parameters. This method is projects. generally applicable to any photogrammetric camera with a narrow field of view, calibrated, stable interior orientation, and Physical Camera Models accurate a priori exterior orientation data. As demonstrated Owing to the dynamic nature of satellite image collection, phoin the paper, for Ikonos satellite imagery, the RPC block togrammetric processing of satellite imagery is more compliadjustment achieves the same accuracy as the ground station cated than is aerial frame camera processing. Aerial cameras block adjustment with the full physical camera model. acquire the entire image at an instant of time with a unique exposure station and orientation. High-resolution pushbroom Background satellite cameras, including Ikonos, use linear sensor arrays

Determining and Visualizing Uncertainty in Estimates of Fiber Orientation From Diffusion Tensor MRI

Abstract: Diffusion tensor MRI (DT-MRI) permits determination of the dominant orientation of structured tissue within an image voxel. This has led to the development of 2D graphical methods for representing fiber orientation and DT-MRI "tractography," which aims to reconstruct the 3D trajectories of white matter fasciculi. Most contemporary fiber orientation mapping schemes and tractography algorithms employ the directional information contained in the eigenvectors of the diffusion tensor to approximate white matter fiber orientation. However, while the uncertainty associated with every estimate of an eigenvector has long been recognized, no attempts to quantify this uncertainty in vivo have been reported. Here, a method is proposed for determining confidence intervals in fiber orientation from real DT-MRI data using the bootstrap method. This is used to construct maps of the "cone of uncertainty," allowing simultaneous viewing of fiber orientation and its uncertainty, and to examine the relationship between orientation uncertainty and tissue anisotropy.

3d modeling system

Abstract: A computer-based 3D modeling system for constructing a virtual 3D representation from a plurality of data images of 2D cross sections having a mutual spatial relationship. The plurality of data images and the associated orientation and positioning information are extractable from a data source module (22). A frame creation module constructs a rectangular frame for each image slice. A texture-mapping module (26) maps the image slice onto the associated frame as a texture. A rotation transform module (28) rotates each frame appropriately about one or more axes based upon the orientation information associated with each data image to achieve the correct orientation in 3D space. A translation transform module (30) translates each frame based upon the positioning information associated with each data image to achieve the correct position in 3D space.

Exact Polyhedral Visual Hulls

Abstract: We propose an exact method for efficiently and robustly computing the visual hull of an object from image contours. Unlike most existing approaches, ours computes an exact description of the visual hull polyhedron associated to polygonal image contours. Furthermore, the proposed approach is fast and allows real-time recovery of both manifold and watertight visual hull polyhedra. The process involves three main steps. First, a coarse geometrical approximation of the visual hull is computed by retrieving its viewing edges, an unconnected subset of the wanted mesh. Then, local orientation and connectivity rules are used to walk along the relevant viewing cone intersection boundaries, so as to iteratively generate the missing surface points and connections. A final connection walkthrough allows us to identify the planar contours for each face of the polyhedron. Implementation details and results with synthetic and real data are presented.

Vision and inertial sensor cooperation using gravity as a vertical reference

Abstract: This paper explores the combination of inertial sensor data with vision. Visual and inertial sensing are two sensory modalities that can be explored to give robust solutions on image segmentation and recovery of 3D structure from images, increasing the capabilities of autonomous robots and enlarging the application potential of vision systems. In biological systems, the information provided by the vestibular system is fused at a very early processing stage with vision, playing a key role on the execution of visual movements such as gaze holding and tracking, and the visual cues aid the spatial orientation and body equilibrium. In this paper, we set a framework for using inertial sensor data in vision systems, and describe some results obtained. The unit sphere projection camera model is used, providing a simple model for inertial data integration. Using the vertical reference provided by the inertial sensors, the image horizon line can be determined. Using just one vanishing point and the vertical, we can recover the camera's focal distance and provide an external bearing for the system's navigation frame of reference. Knowing the geometry of a stereo rig and its pose from the inertial sensors, the collineations of level planes can be recovered, providing enough restrictions to segment and reconstruct vertical features and leveled planar patches.

Method and system for performing surveillance

Abstract: A method for detecting moving objects (fig. 10) and controlling a surveillance system includes a processing module (1016) adapted to receive image information from at least one imaging sensor (104). The system performs motion detection analysis upon captured images and controls the camera (104) in a specific manner upon detection of a moving object. The image processing using the camera (104) orientation, a moving object position, latitude, longitude and altitude within a surveillance area to facilitate mapping images captured by the camera (104) to a reference map of the surveillance area.

Method and apparatus for performing 2D to 3D registration

Abstract: A method and apparatus for performing 2D to 3D registration includes an initialization step and a refinement step. The initialization step is directed to identifying an orientation and a position by knowing orientation information where data images are captured and by identifying centers of relevant bodies. The refinement step uses normalized mutual information and pattern intensity algorithms to register the 2D image to the 3D volume.

Method and system for registering a medical situation associated with a first coordinate system, in a second coordinate system using an MPS system

Abstract: System for registering a first image with a second image, the system including a first medical positioning system for detecting a first position and orientation of the body of a patient, a second medical positioning system for detecting a second position and orientation of the body, and a registering module coupled with a second imager and with the second medical positioning system, the first medical positioning system being associated with and coupled with a first imager, the first imager acquiring the first image from the body, the first imager producing the first image by associating the first image with the first position and orientation, the second medical positioning system being associated with and coupled with the second imager, the second imager acquiring the second image and associating the second image with the second position and orientation, the registering module registering the first image with the second image, according to the first position and orientation and the second position and orientation

Correlating captured images and timed 3D event data

Abstract: A method for correlating tracking data associated with an activity occurring in a three-dimensional space with images captured within the space comprises the steps of: (a) locating a camera with respect to the three-dimensional space, wherein the camera at a given location has a determinable orientation and field of view that encompasses at least a portion of the space; (b) capturing a plurality of images with the camera and storing data corresponding to the images, including a capture time for each image; (c) capturing tracking data from identification tags attached to the people and/or objects within the space and storing the tracking data, including a tag capture time for each time that a tag is remotely accessed; (d) correlating each image and the tracking data by interrelating tracking data having a tag capture time in substantial correspondence with the capture time of each image, thereby generating track data corresponding to each image; (e) utilizing the track data to determine positions of the people and/or objects within the three dimensional space at the capture time of each image; and (f) utilizing the location and orientation of the camera to determine the portion of the space captured in each image and thereby reduce the track data to a track data subset corresponding to people and/or objects positioned within the portion of space captured in each image.

Mosaicing new views: the Crossed-Slits projection

Abstract: We introduce anew kind of mosaicing, where the position of the sampling strip varies as a function of the input camera location. The new images that are generated this way correspond to a new projection model defined by two slits, termed here the Crossed-Slits (X-Slits) projection. In this projection model, every 3D point is projected by a ray defined as the line that passes through that point and intersects the two slits. The intersection of the projection rays with the imaging surface defines the image. X-Slits mosaicing provides two benefits. First, the generated mosaics are closer to perspective images than traditional pushbroom mosaics. Second, by simple manipulations of the strip sampling function, we can change the location of one of the virtual slits, providing a virtual walkthrough of a X-Slits camera; all this can be done without recovering any 3D geometry and without calibration. A number of examples where we translate the virtual camera and change its orientation are given; the examples demonstrate realistic changes in parallax, reflections, and occlusions.

Image processor and image processing method

Abstract: A pattern signal is steganographically encoded in a content object to aid in identifying distortion (e.g., image rotation) to which the object may be subjected. The signal may include a feature that can be used to unambiguously identify correct orientation.

Augmented virtual environments

Abstract: Systems and techniques to implement augmented virtual environments. In one implementation, the technique includes: generating a three dimensional (3D) model of an environment from range sensor information representing a height field for the environment, tracking orientation information of image sensors in the environment with respect to the 3D model in real-time, projecting real-time video from the image sensors onto the 3D model based on the tracked orientation information, and visualizing the 3D model with the projected real-time video. Generating the 3D model can involve parametric fitting of geometric primitives to the range sensor information. The technique can also include: identifying in real time a region in motion with respect to a background image in real-time video, the background image being a single distribution background dynamically modeled from a time average of the real-time video, and placing a surface that corresponds to the moving region in the 3D model.

Texture Segmentation by Multiscale Aggregation of Filter Responses and Shape Elements

Abstract: Texture segmentation is a difficult problem, as is apparentfrom camouflage pictures. A Textured region can containtexture elements of various sizes, each of which can itselfbe textured. We approach this problem using a bottom-upaggregation framework that combines structural characteristicsof texture elements with filter responses. Our processadaptively identifies the shape of texture elements and characterizethem by their size, aspect ratio, orientation, brightness,etc., and then uses various statistics of these propertiesto distinguish between different textures. At the sametime our process uses the statistics of filter responses tocharacterize textures. In our process the shape measuresand the filter responses crosstalk extensively. In addition,a top-down cleaning process is applied to avoid mixing thestatistics of neighboring segments. We tested our algorithmon real images and demonstrate that it can accurately segmentregions that contain challenging textures.

Near-optimal detection of geometric objects by fast multiscale methods

Abstract: We construct detectors for "geometric" objects in noisy data. Examples include a detector for presence of a line segment of unknown length, position, and orientation in two-dimensional image data with additive white Gaussian noise. We focus on the following two issues. i) The optimal detection threshold-i.e., the signal strength below which no method of detection can be successful for large dataset size n. ii) The optimal computational complexity of a near-optimal detector, i.e., the complexity required to detect signals slightly exceeding the detection threshold. We describe a general approach to such problems which covers several classes of geometrically defined signals; for example, with one-dimensional data, signals having elevated mean on an interval, and, in d-dimensional data, signals with elevated mean on a rectangle, a ball, or an ellipsoid. In all these problems, we show that a naive or straightforward approach leads to detector thresholds and algorithms which are asymptotically far away from optimal. At the same time, a multiscale geometric analysis of these classes of objects allows us to derive asymptotically optimal detection thresholds and fast algorithms for near-optimal detectors.

Texture segmentation by multiscale aggregation of filter responses and shape elements

Abstract: Texture segmentation is a difficult problem, as is apparent from camouflage pictures. A textured region can contain texture elements of various sizes, each of which can itself be textured. We approach this problem using a bottom-up aggregation framework that combines structural characteristics of texture elements with filter responses. Our process adaptively identifies the shape of texture elements and characterize them by their size, aspect ratio, orientation, brightness, etc., and then uses various statistics of these properties to distinguish between different textures. At the same time our process uses the statistics of filter responses to characterize textures. In our process the shape measures and the filter responses crosstalk extensively. In addition, a top-down cleaning process is applied to avoid mixing the statistics of neighboring segments. We tested our algorithm on real images and demonstrate that it can accurately segment regions that contain challenging textures.

Shape and motion under varying illumination: unifying structure from motion, photometric stereo, and multiview stereo

Abstract: We present an algorithm for computing optical flow, shape, motion, lighting, and albedo from an image sequence of a rigidly-moving Lambertian object under distant illumination. The problem is formulated in a manner that subsumes structure from motion, multiview stereo, and photometric stereo as special cases. The algorithm utilizes both spatial and temporal intensity variation as cues: the former constrains flow and the latter constrains surface orientation; combining both cues enables dense reconstruction of both textured and textureless surfaces. The algorithm works by iteratively estimating affine camera parameters, illumination, shape, and albedo in an alternating fashion. Results are demonstrated on videos of hand-held objects moving in front of a fixed light and camera.

Modeling and video projection for augmented virtual environments

Abstract: Systems and techniques to implement augmented virtual environments. In one implementation, the technique includes: generating a three dimensional (3D) model of an environment from range sensor information representing a height field for the environment, tracking orientation information of image sensors in the environment with respect to the 3D model in real-time, projecting real-time video from the image sensors onto the 3D model based on the tracked orientation information, and visualizing the 3D model with the projected real-time video. Generating the 3D model can involve parametric fitting of geometric primitives to the range sensor information. The technique can also include: identifying in real time a region in motion with respect to a background image in real-time video, the background image being a single distribution background dynamically modeled from a time average of the real-time video, and placing a surface that corresponds to the moving region in the 3D model.

Augmented reality system and method

Abstract: An augmented reality system comprising: a camera (19) for capturing an image, the camera being movably locatee at a local site, a registering unit (9), generating graphics and registering the generated graphics to the image from the camera, to provide a composite augmented reality image, a display device (5) located at a remote site, physically separated from the local site, for displaying a view comprising the composite augmented reality image, and a communication link (1), for communication of information between the local and the remote site, and a specifying unit (7), for specification of a position and an orientation in the remote site. The registering unit is adapted for registering the generated graphical representation to the image in dependence of the specified position and orientation, and the camera is arranged such that its position and orientation is dependent on the specified position and orientation.

A 128 /spl times/ 128 pixel 120 dB dynamic range vision sensor chip for image contrast and orientation extraction

Abstract: A vision sensor for low-cost, fast, and robust vision systems is described. The sensor includes an on-chip analog computation of contrast magnitude and direction of image features. A temporal ordering of this information according to the contrast magnitude is used to reduce the amount of data delivered. This sensor, realized in a 0.5-/spl mu/m two-poly three-metal technology, features a contrast sensitivity of 2%, a contrast direction precision of /spl plusmn/3/spl deg/, and an illumination dynamic range of 120 dB. Applications with uncontrolled lighting conditions are ideal for this sensor.

Detection and recognition of license plate characters with different appearances

Abstract: This paper proposes an approach to developing an automatic license plate recognition system Car images are taken from various positions outdoors Because of the variations of angles from the camera to the car, license plates have various locations and rotation angles in an image In the license plate detection phase, the magnitude of the vertical gradients is used to detect candidate license plate regions These candidate regions are then evaluated based on three geometrical features: the ratio of width and height, the size and the orientation The last feature is defined by the major axis In the character recognition phase, we must detect character features that are non-sensitive to the rotation variations The various rotated character images of a specific character can be normalized to the same orientation based on the major axis of the character image The crossing counts and peripheral background area of an input character image are selected as the features for rotation-free character recognition Experimental results show that the license plates detection method can correctly extract all license plates from 102 car images taken outdoors and the rotation-free character recognition method can achieve an accuracy rate of 986%

Manhattan world: orientation and outlier detection by Bayesian inference

Abstract: This letter argues that many visual scenes are based on a "Manhattan" three-dimensional grid that imposes regularities on the image statistics. We construct a Bayesian model that implements this assumption and estimates the viewer orientation relative to the Manhattan grid. For many images, these estimates are good approximations to the viewer orientation (as estimated manually by the authors). These estimates also make it easy to detect outlier structures that are unaligned to the grid. To determine the applicability of the Manhattan world model, we implement a null hypothesis model that assumes that the image statistics are independent of any three-dimensional scene structure. We then use the log-likelihood ratio test to determine whether an image satisfies the Manhattan world assumption. Our results show that if an image is estimated to be Manhattan, then the Bayesian model's estimates of viewer direction are almost always accurate (according to our manual estimates), and vice versa.

System and method for inserting content into an image sequence

Abstract: A real-time system and method for inserting perspective correct content into an image sequence are presented. The invention inserts the content with the location, size, orientation, shape and occlusion properties that are appropriate for the camera view represented by the image sequence. Both static and dynamic content insert positions are supported. The location, size, orientation and shape of the inserted content are determined independently of the image sequence content. Furthermore, no knowledge of three dimensional real world space locations or real world measurements, as related to the content of the image sequence, is used during the content insert process.