Pascal Monasse

Gustave Eiffel University · Institut d'électronique et d’'informatique Gaspard Monge (IGM)

Combining image and LiDAR draws increasing interest in surface reconstruction, city and building modeling for constructing 3D virtual reality models because of their complementary nature. However, to gain from this complementarity, these data sources must be precisely registered. The objective of this study is to propose a new primitive based regis...

Scientific research requires that results can be accessed, tested, and replicated by other researchers. In general, performing reproducible research is not simple and it can be even impossible (for example, in the case of astrophysics replicating singular events if even out of the control of the researcher). However, in computational sciences there...

The registration of indoor and outdoor scans with a precision reaching the level of geometric noise represents a major challenge for Indoor/Outdoor building modeling. The basic idea of the contribution presented in this paper consists in extracting planar polygons from indoor and outdoor LiDAR scans, and then matching them. In order to cope with th...

Deep multi-view stereo (deep MVS) methods have been developed and extensively compared on simple datasets, where they now outperform classical approaches. In this paper, we ask whether the conclusions reached in controlled scenarios are still valid when working with Internet photo collections. We propose a methodology for evaluation and explore the...

This book constitutes the thoroughly refereed post-workshop proceedings of the Third International Workshop on Reproducible Research in Pattern Recognition, RRPR 2021, held as a virtual event, in January 2021. The 8 revised full papers, presented together with 6 short papers, were carefully reviewed and selected from 18 submissions. The papers were...

We tackle the problem of finding accurate and robust keypoint correspondences between images. We propose a learning-based approach to guide local feature matches via a learned approximate image matching. Our approach can boost the results of SIFT to a level similar to state-of-the-art deep descriptors, such as Superpoint, ContextDesc, or D2-Net and...

Detecting planar structures in point clouds is a very central step of the point cloud processing pipeline as many Lidar scans, in particular in anthropic environments, present such planar structures. Many improvements have been proposed to RANSAC and the Hough transform, the two major types of plane detection methods. An important limitation howeve...

We propose an algorithm computing the tree of shapes of an image, a unified variation of the component trees, proceeding from the root to the leaf shapes in a recursive fashion. It proceeds differently from existing algorithms that start from leaves, which are regional extrema of intensity, and build the intermediate shapes up to the root, which is...

In engineering sciences, parameter estimation is a challenging problem consisting in computing the parameters of a parametric model that fit observed data. The system is defined by unknown parameters and sometimes internal constraints. The observed data provide constraints on the parameters. This problem is particularly difficult when some observat...

The parameter estimation problem is a challenging problem in engineering sciences consisting in computing the parameters of a parametric model that fit observed data. The system is defined by unknown parameters and sometimes internal constraints. The observed data provide constraints on the parameters. The parameter estimation problem is particular...

We explain how the B-spline interpolation of signals and, in particular, of images can be ef-ficiently performed by linear filtering. Based on the seminal two-step method proposed byUnser et al. in 1991, we propose two slightly different prefiltering algorithms whose precisionsare proven to be controlled thanks to a rigorous boundary handling. This...

We explore the advantages offered by the trifocal tensor in the pose estimation of a triplet of cameras as opposed to computing the relative poses pair by pair with the fundamental matrix. Theoretically, the trinilearities characterize uniquely three corresponding image points in a tighter way than the three epipolar equations and this translates i...

We identified design problems related to the architecture, ergonomy, and performance in the previous version of the Image Processing on Line (IPOL) demonstration system. In order to correct them we moved to an architecture of microservices and performed many refac-torings. This article first describes the state of the art in Reproducible Research p...

The aim of a Super resolution (SR) technique is to construct a high-resolution image from a sequence of observed low-resolution ones of the same scene. One major roadblock of an SR reconstitution is removing noise and blur without destroying edges. We propose a novel multiframe image SR algorithm based on a convex combination of Bilateral Total Var...

We detail in this paper the numerical implementation of the so-called image curvature microscope, an algorithm that computes accurate image curvatures at subpixel resolution, and yields a curvature map conforming with our visual perception. In contrast to standard methods, which would compute the curvature by a finite difference scheme, the curvatu...

The OpenMVG C++ library provides a vast collection of multiple-view geometry tools and algorithms to spread the usage of computer vision and structure-from-motion techniques. Close to the state-of-the-art in its domain, it provides an easy access to common tools used in 3D reconstruction from images. Following the credo “Keep it simple, keep it mai...

We propose a multiscale extension of a well-known line segment detector, LSD. We show that its multiscale nature makes it much less susceptible to over-segmentation and more robust to low contrast and less sensitive to noise, while keeping the parameter-less advantage of LSD and still being fast. We also present here a dense gradient filter that di...

Usual Structure-from-Motion (SfM) techniques require at least trifocal overlaps to calibrate cameras and reconstruct a scene. We consider here scenarios of reduced image sets with little overlap, possibly as low as two images at most seeing the same part of the scene. We propose a new method, based on line coplanarity hypotheses, for estimating the...

This paper addresses the question of identifying the right camera direct or inverse distortion model permitting a high subpixel precision fit to real camera distortion. Five classic camera distortion models are reviewed and their precision compared for direct or inverse distortion. By definition, the three radially symmetric models can only model a...

This book constitutes the thoroughly refereed post-proceedings of the First International Workshop on Reproducible Research in Pattern Recognition, RRPR 2016, held in Cancún, Mexico, in December 2016. The 12 revised full papers, among them 2 invited talks, presented were carefully reviewed and selected from 16 submissions. They focus on pattern rec...

Usual Structure from Motion techniques based on feature points have a hard time on scenes with little texture or presenting a single plane, as in indoor environments. Line segments are more robust features in this case. We propose a novel geometrical criterion for two-view pose estimation using lines, that does not assume a Manhattan world. We also...

In a stereo image pair, the fundamental matrix encodes the rigidity constraint of the scene. It combines the internal parameters of both cameras (which can be the same) and their relative position and orientation. It associates to image points in one view the so-called epipolar line in the other view, which is the locus of projection of the same 3D...

A method to extract digital elevation models of the ground from the images of an airborne Fourier transform spectrometer is presented. Such information allows to construct accurate interferograms in non-flat areas, where they are otherwise flawed.

After four years of development of the Image Processing On Line journal (IPOL), this article presents a first analysis and overview of its scientific and technical development. The main issues met and overcome from the beginning of the journal are described with a focus on the purpose of the journal to establish a state of the art on the main Image...

Visual correspondence is the key for 3D reconstruction in binocular stereovision. Local methods perform block-matching to compute the disparity, or apparent motion, of pixels between images. The simplest approach computes the distance of patches, usually square windows, and assumes that all pixels in the patch have the same disparity. A prominent a...

This paper introduces a topological approach to local invariant feature detection motivated by Morse theory. We use the critical points of the graph of the intensity image, revealing directly the topology information as initial "interest" points. Critical points are selected from what we call a treebased shape-space. Specifically, they are selected...

We evaluate and improve the matching precision of the SIFT method [1], defined as the root mean square error (RMSE) under a ground truth geometric transform. We first argue that the matching precision reflects to some extent the average relative localization precision between two images. For scale invariant feature detectors like SIFT, we show that...

Estimating the depth, or equivalently the disparity, of a stereo scene is a challenging problem in computer vision. The method proposed by Rhemann et al. in 2011 is based on a filtering of the cost volume, which gives for each pixel and for each hypothesized disparity a cost derived from pixel-by-pixel comparison. The filtering is performed by the...

Binocular stereovision estimates the three-dimensional shape of a scene from two photographs taken from different points of view. In rectified epipolar geometry, this is equivalent to a matching problem. This article describes a method proposed by Kolmogorov and Zabih in 2001, which puts forward an energy-based formulation. The aim is to minimize a...

We present an approach to enhance the accuracy of structure from motion (SfM) in the two-view case. We first answer the question: "fewer data with higher accuracy, or more data with less accuracy?" For this, we establish a relation between SfM errors and a function of the number of matches and their epipolar errors. Using an accuracy estima-tor of...

Reflective symmetry can be used as a strong prior for many computer vision tasks. We interpret the planar reflective symmetry detection by using the property of an involution, which implies that two pairs of matched points are enough to define a planar reflective symmetry observed from a non-frontal viewpoint. This makes the reflective symmetry est...

We achieve a precise camera calibration using circular control points by, first, separation of the lens distortion parameters from other camera parameters and computation of the distortion field in advance by using a calibration harp. Second, in order to compensate for perspective bias, which is prone to occur when using a circled pattern, we incor...

This paper presents a review, analysis, and comparison of numerical methods implementing the curvature motion and the affine curvature motion for two-dimensional (2D) images, shapes, and curves. These curvature scale spaces allow one, in principle, to compute an accurate multiscale curvature in digital images. The fastest and most invariant of them...

The fundamental matrix is a two-view tensor that plays a central role in Computer Vision geometry. We address its robust estimation given correspondences between image features. We use a non-parametric estimate of the distribution of image features, and then follow a probabilistic approach to select the best possible set of inliers among the given...

The normalized eight-point algorithm is broadly used for the computation of the fundamental matrix between two images given a set of correspondences. However, it performs poorly for low-size datasets due to the way in which the rank-two constraint is imposed on the fundamental matrix. We propose two new algorithms to enforce the rank-two constraint...

Multi-view structure from motion (SfM) estimates the position and orientation of pictures in a common 3D coordinate frame. When views are treated incrementally, this external calibration can be subject to drift, contrary to global methods that distribute residual errors evenly. We propose a new global calibration approach based on the fusion of rel...

In a multiple-view image acquisition process, color consistency is not ensured. This is an important problem for image fusion tasks: object texturing or mosaics blending for example. In automatic mode, the camera adapts its settings –shutter-speed and aperture– to the captured image content. Therefore the color of objects changes over an image sequ...

This paper addresses the problem of lateral chromatic aberration correction in images through color planes warping. We aim at high precision (largely sub-pixel) realignment of color channels. This is achieved thanks to two ingredients: high precision keypoint detection, which in our case are disk centers, and more general correction model than what...

L'estimation de modèle consiste à identifier un modèle parmi des données bruitées. Ce problème n'est pas trivial et l'état de l'art présente de nombreuses solutions pour résoudre ce problème. Le plus souvent les solutions max-consensus ou RANSAC sont utilisées. Ces solutions proposent de rechercher par tirages aléatoires plusieurs solutions et de co...

We present an efficient algorithm to fuse two-view correspondences into multi-view consistent tracks. The proposed method relies on the Union-Find [1] algorithm to solve the fusion problem. It is very simple and has a lower computational complexity than other available methods. Our experiments show that it is faster and computes more tracks.

Structure from Motion (SfM) algorithms take as input multi-view stereo images (along with internal calibration information) and yield a 3D point cloud and camera orientations/poses in a common 3D coordinate system. In the case of an incremental SfM pipeline, the process requires repeated model estimations based on detected feature points: homograph...

This paper introduces the concepts of "self-consistency" and "universality" to evaluate high precision camera distortion models. Self-consistency is evaluated by the residual error when the distortion generated with a certain model is corrected by the best parameters for the same model (used in reverse way, which is common practice). Analogously, u...

Structure from Motion (SfM) algorithms take as input multi-view stereo images (along with internal calibration information) and yield a 3D point cloud and camera orientations/poses in a common 3D coordinate system. In the case of an incremental SfM pipeline, the process requires repeated model estimations based on detected feature points: homograph...

This paper addresses the high-precision measurement of the distortion of a digital camera from photographs. Traditionally, this distortion is measured from photographs of a flat pattern that contains aligned elements. Nevertheless, it is nearly impossible to fabricate a very flat pattern and to validate its flatness. This fact limits the attainable...

The RANSAC algorithm (RANdom SAmple Consensus) is a robust method to estimate parameters of a model fitting the data, in presence of outliers among the data. Its random nature is due only to complexity considerations. It iteratively extracts a random sample out of all data, of minimal size sufficient to estimate the parameters. At each such trial,...

Detecting a pattern in an image by receiving the image of a pattern and storing the image in a memory, where the pattern is composed of shapes that have geometrical properties that are invariant under near projective transforms. In some embodiments the process detects shapes in the image using the geometrical properties of the shapes, determines th...

The standard setup in reconstructing the three-dimensional geometry of a scene from a pair of stereo images is to have them rectified, in which case the apparent motion of points is horizontal. With pinhole cameras, it is always possible to find two homographies that rectify the images. The method of Fusiello and Irsara assumes that both cameras ar...

Plumb line lens distortion correction methods permit to avoid numerical compensation between the camera internal and external parameters in global calibration method. Once the distortion has been corrected by a plumb line method, the camera is ensured to transform, up to the distortion precision, 3D straight lines into 2D straight lines, and theref...

A method and apparatus to extract a dense three-dimensional model of an observed scene or object from one or more sequences of images acquired with an imaging device such as a camera or camcorder, or clusters of cameras and/or camcorders. In some embodiments the method includes capturing an image sequence by a camera moving with a translateral moti...

This paper presents an image processing algorithm simulating a sub-pixel evolution of an image by mean curvature motion or by affine curvature motion. The sub-pixel algorithm computes the image curvature directly on the smoothed level lines, and yields a microscopic visualization of the curvature map revealing many image details, and getting rid of...

Chapter 6 proposed an algorithm computing the tree of shapes based on an interpretation of the discrete image as a piecewise constant bivariate function. The corresponding interpolation operator, commonly known as nearest neighbor or order 0 interpolation, does not removes the pixelization effect. Bilinear or order 1 interpolation is notoriously pr...

Chapter 2 started by presenting the tree of shapes as a fusion of the tree of connected components of upper level sets and the tree of connected components of lower level sets, known commonly as the component trees. We come back in this chapter to a constructive study of this fusion, leading to an algorithm applicable in any dimension. Direct and m...

We gather here some applications of the FLST and the tree decomposition to image processing. All were published in the literature, and although they do not pretend to represent the state of the art in their respective domain, we hope they emphasize its versatility and convenience. In several instances, level lines provide a superior alternative for...

This paper points out and attempts to remedy a serious discrepancy in results obtained by global calibration methods: The re-projection error can be rendered very small by these methods, but we show that the optical distortion correction is far less accurate. This discrepancy can only be explained by internal error compensations in the global metho...

Image stereo-rectification is the process by which two images of the same solid scene undergo homographic transforms, so that their corresponding epipolar lines coincide and become parallel to the x-axis of image. A pair of stereo-rectified images is helpful for dense stereo matching algorithms. It restricts the search domain for each match to a li...

The tree of shapes of an image, as a unified mix of the component trees, was presented and analyzed in the present notes. Its existence was proven for a semicontinuous image. It relies on the simple model according to which shapes have mostly uniform reflectance properties and are distinguishable from the background. This model has no parameters an...

We present here a first algorithm, the Fast Level Sets Transform, that computes the tree of shapes from an input image. Although it is applicable in any dimension, its efficiency relies on specific properties of a dimension 2 domain. Although a direct derivation of the existence of the tree of shapes in the discrete case would be possible, a judici...

The aim ofMorse theory is to describe the topological changes of the (iso)level sets of a real valued function in terms of its critical points. Our purpose in this chapter is to describe several different notions of critical values (one of them will be called critical value, the other singular value) and prove that they are equivalent. One of those...

This chapter is intended as a gentle and casual introduction to the themes developed in these lecture notes. It cites only one reference book for each theme and omits voluntarily all further citations to let it remain an overview of usually independent but here connected topics: Morse theory, topography and mathematical morphology. The three have i...

Reprojection of texture issued from cameras on a mesh estimated from multi-view reconstruction is often the last stage of the pipeline, used for rendering, visualization, or simulation of new views. Errors or imprecisions in the recovered 3D geometry are particularly noticeable at this stage. Nevertheless, it is sometimes desirable to get a visuall...

In this paper, we propose a novel method to simultaneously and accurately estimate the 3D shape and 3D motion of a dynamic scene from multiple-viewpoint calibrated videos. We follow a variational approach in the vein of previous work on stereo reconstruction and scene flow estimation. We adopt a representation of a dynamic scene by an animated mesh...

Methods and systems for performing object directed recognition based on two-dimensional images and three-dimensional models. Transforms are used to map features of an object to an image seen from a vantage point and to map features of images best seen from the vantage points to three-dimensional models. Mapping of features and images to the three-d...

In this report, we describe a novel method to simultaneously and accurately estimate the 3D shape and 3D motion of a dynamic scene from multiple-viewpoint calibrated videos. We follow a variational approach in the vein of previous work on stereo reconstruction and scene flow estimation. We adopt a representation of a dynamic scene by an animated me...

The tree of shapes of an image is an ordered structure which permits an efficient manipulation of the level sets of an image, modeled as a real continuous function defined on a rectangle of \mathbbRN{\mathbb{R}}^N , N≥2. In this paper we construct the tree of shapes of an image by fusing both trees of connected components of upper and lower level...

We propose a novel method for estimating super-dense 3-D geometry of urban scenes and other terrains from an image sequence captured by a calibrated airborne video camera. The flight is assumed to be at constant velocity in a piecewise straight path and the camera is oriented at nadir. Under these conditions, the epipolar lines are stationary. We e...

This paper presents a software-based parameter-free method for the demultiplexing of