International audience

Visual place recognition is a key to unlocking spatial navigation for animals, humans and robots. While state-of-the-art approaches are trained in a supervised manner and therefore hardly capture the information needed for generalizing to unusual conditions, we argue that self-supervised learning may help abstracting the place representation so that it can be foreseen, irrespective of the conditions. More precisely, in this paper, we investigate learning features that are robust to appearance modifications while sensitive to geometric transformations in a self-supervised manner. This dual-purpose training is made possible by combining the two self-supervision main paradigms, i.e. contrastive and predictive learning. Our results on standard benchmarks reveal that jointly learning such appearance-robust and geometry-sensitive image descriptors leads to competitive visual place recognition results across adverse seasonal and illumination conditions, without requiring any humanannotated labels.1.

To automatically localize a target object in an image is crucial for many computer vision applications. To represent the 2D object, ellipse labels have recently been identified as a promising alternative to axis-aligned bounding boxes. This paper further considers 3D-aware ellipse labels, i.e., ellipses which are projections of a 3D ellipsoidal approximation of the object, for 2D target localization. Indeed, projected ellipses carry more geometric information about the object geometry and pose (3D awareness) than traditional 3D-agnostic bounding box labels. Moreover, such a generic 3D ellipsoidal model allows for approximating known to coarsely known targets. We then propose to have a new look at ellipse regression and replace the discontinuous geometric ellipse parameters with the parameters of an implicit Gaussian distribution encoding object occupancy in the image. The models are trained to regress the values of this bivariate Gaussian distribution over the image pixels using a statistical loss function. We introduce a novel non-trainable differentiable layer, E-DSNT, to extract the distribution parameters. Also, we describe how to readily generate consistent 3D-aware Gaussian occupancy parameters using only coarse dimensions of the target and relative pose labels. We extend three existing spacecraft pose estimation datasets with 3D-aware Gaussian occupancy labels to validate our hypothesis. Labels and source code are publicly accessible here: https://cvi2.uni.lu/3d-aware-obj-loc/.

While end-to-end approaches have achieved state-of-the-art performance in many perception tasks, they are not yet able to compete with 3D geometry-based methods in pose estimation. Moreover, absolute pose regression has been shown to be more related to image retrieval. As a result, we hypothesize that the statistical features learned by classical Convolutional Neural Networks do not carry enough geometric information to reliably solve this inherently geometric task. In this paper, we demonstrate how a translation and rotation equivariant Convolutional Neural Network directly induces representations of camera motions into the feature space. We then show that this geometric property allows for implicitly augmenting the training data under a whole group of image plane-preserving transformations. Therefore, we argue that directly learning equivariant features is preferable than learning data-intensive intermediate representations. Comprehensive experimental validation demonstrates that our lightweight model outperforms existing ones on standard datasets.1

In computer vision, estimating camera pose from correspondences between 3D geometric entities and their projections into the image is a widely investigated problem. Although most state-of-the-art methods exploit simple primitives such as points or lines, and thus require dense scene models, the emergence of very effective CNN-based object detectors in the recent years have paved the way to the use of much lighter 3D models composed solely of a few semantically relevant features. In that context, we propose a novel model-based camera pose estimation method in which the scene is modeled by a set of virtual ellipsoids. We show that 6-DoF camera pose can be determined by optimizing only the three orientation parameters, and that at least two correspondences between 3D ellipsoids and their 2D projections are necessary in practice. We validate the approach on both simulated and real environments.

We are interested in AR applications which take place in man-made GPS-denied environments, as industrial or indoor scenes. In such environments, relocalization may fail due to repeated patterns and large changes in appearance which occur even for small changes in viewpoint. We investigate in this paper a new method for relocalization which operates at the level of objects and takes advantage of the impressive progress realized in object detection. Recent works have opened the way towards object oriented reconstruction from elliptic approximation of objects detected in images. We go one step further and propose a new method for pose computation based on ellipse/ellipsoid correspondences. We consider in this paper the practical common case where an initial guess of the rotation matrix of the pose is known, for instance with an inertial sensor or from the estimation of orthogonal vanishing points. Our contributions are twofold: we prove that a closed form estimate of the translation can be computed from one ellipse-ellipsoid correspondence. The accuracy of the method is assessed on the LINEMOD database using only one correspondence. Second, we prove the effectiveness of the method on real scenes from a set of object detections generated by YOLO. A robust framework that is able to choose the best set of hypotheses is proposed and is based on an appropriate estimation of the reprojection error of ellipsoids. Globally, considering pose at the level of object allows us to avoid common failures due to repeated structures. In addition, due to the small combinatory induced by object correspondences, our method is well suited to fast rough localization even in large environments.

Given two views of the same scene, usual correspondence geometry estimation techniques exploit the well-established effectiveness of keypoint descriptors. However, such features have a hard time in poorly textured man-made environments, possibly containing repetitive patterns and/or specularities, such as industrial places. In that paper, we propose a novel method for two-view epipolar and planar geometry estimation that first aims at detecting and matching physical vertical planes frequently present in these environments, before estimating corresponding homographies. Inferred local correspondences are finally used to improve fundamental matrix estimation. The gain in precision is demonstrated on industrial and urban environments.