In central catadioptric systems lines in a scene are projected to conic curves the image. This work studies geometry of projection and its use calibration. It is shown that where mapped possess several projective invariant properties. From these properties, it follows any system can be fully calibrated from an image three or more lines. The absolute conic, relative pose between camera mirror, shape reflective surface recovered using geometric construction based on loci projected. result...

This paper presents a new algorithm for the extrinsic calibration of perspective camera and an invisible 2D laser-rangefinder (LRF). The is achieved by freely moving checkerboard pattern in order to obtain plane poses coordinates depth readings LRF reference frame. problem estimating rigid displacement between two sensors formulated as one registering set planes lines 3D space. It proven first time that alignment three plane-line correspondences has at most eight solutions can be determined...

An imaging system with a single effective viewpoint is called central projection system. The conventional perspective camera an example of Systems using mirrors to enhance the field view while keeping unique center are also examples systems. Perspective image formation can be described by linear model well known properties. In general catadioptric imaging, mapping between points in world and highly nonlinear. paper establishes for made up three functions: function into oriented projective...

Keypoint detection and matching is of fundamental importance for many applications in computer robot vision. The association points across different views problematic because image features can undergo significant changes appearance. Unfortunately, state-of-the-art methods, like the scale-invariant feature transform (SIFT), are not resilient to radial distortion that often arises images acquired by cameras with microlenses and/or wide field-of-view. This paper proposes modifications SIFT...

Medical endoscopy is used in a wide variety of diagnostic and surgical procedures. These procedures are renowned for the difficulty orienting camera instruments inside human body cavities. The small size lens causes radial distortion image, which hinders navigation process leads to errors depth perception object morphology. This article presents complete software-based system calibrate correct clinical real time. Our can be with any type medical endoscopic technology, including...

When deploying a heterogeneous camera network or when we use cheap zoom cameras like in cell-phones, it is not practical, if impossible to off-line calibrate the radial distortion of each using reference objects. It rather desirable have an automatic procedure without strong assumptions about scene. In this paper, present new algorithm for estimating epipolar geometry two views where can be radially distorted with different factors. first literature solving case left and right view linearly...

The article describes a new algorithm for calibrating Kinect sensor that achieves high accuracy using only 6 to 10 image-disparity pairs of planar checkerboard pattern. method estimates the projection parameters both color and depth cameras, relative pose between them, function converts kinect disparity units (kdu) into metric depth. We build on recent work Herrera et. al [8] uses large number input frames multiple iterative minimization steps obtaining very accurate calibration results....

Affine Correspondences (ACs) are more informative than Point (PCs) that used as input in mainstream algorithms for Structure-from-Motion (SfM). Since ACs enable to estimate models from fewer correspondences, its use can dramatically reduce the number of combinations during iterative step sample-and-test exists most SfM pipelines. However, using instead PCs passes by fully understanding relations between and multi-view geometry, well establishing practical, effective AC-based algorithms. This...

The paper proposes a new calibration algorithm for cameras with lens distortion, that uses single image of planar chessboard pattern acquired in general position.The radial distortion is modeled using the first order division model, and method provides closed form estimation intrinsic parameters coefficient.The experimental evaluation shows accuracy comparable to state-of-the-art algorithms requiring multiple input images.We believe our approach particularly well suited medical endoscopes...

This article presents a new global approach for detecting vanishing points and groups of mutually orthogonal directions using lines detected in images man-made environments. These two multi-model fitting problems are respectively cast as Uncapacited Facility Location (UFL) Hierarchical (HFL) instances that efficiently solved message passing inference algorithm. We also propose functions measuring the consistency between an edge putative point, computing point defined by subset edges....

In this paper we propose a hand gesture recognition system that relies on color and depth images, small pose sensor the human palm. Monocular stereo vision systems have been used for recognition, but with limited scope due to limitations texture, illumination, etc. New RGB-Depth sensors, reply projected light such as Microsoft Kinect, overcome many of those limitations. However, point clouds gestures are still in cases noisy partially occluded, is not trivial. Hand much more complex than...

Global 3D point cloud registration has been solved by finding putative matches between the clouds for establishing alignment hypotheses. A naive approach would try to perform exhaustive search of triplets with a cubic runtime complexity in number data points. Super4PCS reduces this linear making use sets 4 coplanar This paper proposes 2-Point-Normal Sets (2PNS), new global that advances using 2 points and their normals generating The dramatic improvement 2PNS when compared is demonstrated...

Odometry consists in using data from a moving sensor to estimate change position over time. It is crucial step for several applications robotics and computer vision. This paper presents novel approach estimating the relative motion between successive RGB-D frames that uses plane-primitives instead of point features. The planes scene are extracted estimation cast as plane-to-plane registration problem with closed-form solution. Point features only cases where plane surface configuration...

The article concerns the automatic calibration of a camera with radial distortion from single image. It is known that, under mild assumption square pixels and zero skew, lines in scene project into circles image, three suffice to calibrate up an ambiguity between focal length distortion. results highly depend on accurate circle estimation, which hard accomplish because tend short circular arcs. To overcome this problem, we show given arc edge, it possible robustly determine line that goes...

This paper describes a machine learning approach for visual object detection and recognition which is capable of processing images rapidly achieving high rates. framework demonstrated on, in part motivated by, the task human-robot interaction. There are three main parts on this framework. The first person's face used as preprocessing system to second stage person interacting with robot, third one hand detection. technique based Haar-like features introduced by Viola et al. then improved...

Endoscopic camera for surgical navigation and 3D visualization requires precise stable estimates of the calibration parameters. The estimation hand-eye transform between frame opto-tracked body endoscope is an important issue calibration. This paper presents a new method problem. most popular directly in special euclidean group SE(3) by computing separately rotation translation. second famous approach formulates problem dual quaternion space jointly In first glance, simultaneous seems to be...

Estimating the amount and center of distortion from lines in scene has been addressed literature by so-called ``plumb-line'' approach. In this paper we propose a new geometric method to estimate not only parameters but entire camera calibration (up an ``angular'' scale factor) using minimum 3 lines. We framework for unsupervised simultaneous detection natural image estimation, enabling robust single image. Comparative experiments with existing automatic approaches estimation ground truth...

Knee arthritis is a common joint disease that usually requires total knee arthroplasty. There are multiple surgical variables have direct impact on the correct positioning of implants, and an optimal combination all these most challenging aspect procedure. Usually, preoperative planning using computed tomography scan or magnetic resonance imaging helps surgeon in deciding suitable resections to be made. This work proof concept for navigation system supports following plan. Existing solutions...