Affine Projection Model Research Articles

Automatic Tool Landmark Detection for Stereo Vision in Robot-Assisted Retinal Surgery

Computer vision and robotics are being increasingly applied in medical interventions. Especially in interventions where extreme precision is required, they could make a difference. One such application is robot-assisted retinal microsurgery. In recent works, such interventions are conducted under a stereo-microscope, and with a robot-controlled surgical tool. The complementarity of computer vision and robotics has, however, not yet been fully exploited. In order to improve the robot control, we are interested in three-dimensional (3-D) reconstruction of the anatomy and in automatic tool localization using a stereo microscope. In this letter, we solve this problem for the first time using a single pipeline, starting from uncalibrated cameras to reach metric 3-D reconstruction and registration, in retinal microsurgery. The key ingredients of our method are 1) surgical tool landmark detection, and 2) 3-D reconstruction with the stereo microscope, using the detected landmarks. To address the former, we propose a novel deep learning method that detects and recognizes keypoints in high-definition images at higher than real-time speed. We use the detected two-dimensional keypoints along with their corresponding 3-D coordinates obtained from the robot sensors to calibrate the stereo microscope using an affine projection model. We design an online 3-D reconstruction pipeline that makes use of smoothness constraints and performs robot-to-camera registration. The entire pipeline is extensively validated on open-sky porcine eye sequences. Quantitative and qualitative results are presented for all steps.

Binocular Camera Self Calibration Method based on the Affine Projection Model

Binocular Camera Self Calibration Method based on the Affine Projection Model

ORTORETIFICAÇÃO DE IMAGENS DE ALTA RESOLUÇÃO UTILIZANDO OS MODELOS APM (AFFINE PROJECTION MODEL) E RPC (RATIONAL POLYNOMIAL COEFFYCIENT)

Absolute positional precision of high resolution image has been studied recently. The improvement of methods and techniques of extraction of information has been developed focusing mathematical models for ortho-rectification and 3D composition of high resolution images. Among the promising ‘alternative’ sensor orientation models investigated has been an approach based on affine projection. This paper briefly summarizes the theory and validity of the affine model as configured for application to sensor orientation and geopositioning . The results of experiments with one Ikonos Stereo and Quickbird (Standard) has provided a comparison of mathematical model APM (Affine Projection Model) and the Functional Rational model (RFM – Rational Function Model).

A recursive factorization method for the metric affine projection model

AbstractThe task of reconstructing camera motion and the shapes of objects from multiple images is a fundamental and important one in computer vision. Although among the reconstruction methods, the batch processing factorization method proposed by Tomasi and Kanade gives mathematically stable and good reconstruction results, it requires computational time that increases proportionately with the number of images handled and thus cannot be easily applied to real‐time processing. This paper proposes a recursive factorization method for a metric affine projection model, which allows a factorization method to be applied to real‐time processing. The proposed method makes it possible to minimize the sizes of the matrices handled and to shorten the computational time significantly while realizing a reconstruction accuracy comparable to that of motion matrices, by lossless compression of the motion information of the motion matrices by means of principal component analyses. In addition, the proposed method facilitates the processing of occlusion points and the addition of feature points. © 2006 Wiley Periodicals, Inc. Syst Comp Jpn, 37(7): 83–93, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/scj.10134

高解像度ＩＫＯＮＯＳ衛星画像を用いた精密３次元計測

An investigation of the accuracy potential of IKONOS lm satellite imagery is reported. Two sensor orientation/triangulation models, the RPC models with bias compensation and the affine projection model, are applied to stereo- and three-image configurations of Geo imagery with the aim of achieving 3D geopositioning to sub-metre accuracy. Test results from the Melbourne IKONOS Testfield are reported and these show that with modest provision of quality ground control, IKONOS Geo imagery can yield 3D object point determination to an accuracy of 0.5m in planimetry and 0.7m in height. The accuracy achieved is not only consistent with expectations for rigorous sensor orientation models, but is also readily attainable in practice with only a small number of ground control points being required.

Rejection of mismatched correspondences along the affine epipolar line

Identifying the correspondences of points between images is an important task. It finds applications in depth recovery, motion recovery, photogrammetry, and robotics. Commonly, one uses the epipolar geometry to constrain the problem from a two-dimensional to a one-dimensional search. However, the estimation of the epipolar parameters itself depends on finding a sufficiently accurate set of initial point correspondences. Mismatched correspondences present themselves as outliers in this parameter estimation. Even if the epipolar geometry is perfectly found, it is still not sufficient to disambiguate false matches whose points lie along the epipolar line. This paper attempts to identify (and thereby reject) such mismatched correspondences, within the framework of an affine projection model and the use of three images. The robust Least Median of Squares (LMedS) approach is utilized in two stages. First, it is used to estimate the epipolar parameters and reject those point-correspondences whose orthogonal distance to the epipolar lines are significant. Second, points are resolved along their epipolar lines and the LMedS is used to solve for a consistent rotation for all points using three images.