Self-supervised monocular depth prediction is a widely researched field that aims to provide a better scene understanding. However, most existing methods prioritize prediction accuracy over computation cost, which can hinder the deployment of these methods in real-world applications. Our objective is to propose a solution that efficiently compresses the depth map while maintaining a high level of accuracy for navigation purpose. The proposed method is an expansion of the work presented in N-QGN, which utilizes a quadtree representation for compression. This approach has already shown promising results, but we aim to improve it further by making it more accurate, faster, and easier to train. Therefore, we introduce a new method that directly predicts the quadtree structure, resulting in a more consistent prediction, and we revise the network architecture to be lighter and produce state-of-the-art accuracy results, depending on the data compression rate. The new implementation is also faster, making it more suitable for real-time applications. Experiments have been conducted on various scene configuration highlighting the capability of the method to efficiently predicting a reliable quadtree depth representation of the scene at low computation cost and high accuracy.

Monocular depth estimation has been a popu-lar area of research for several years, especially since self-supervised networks have shown increasingly good results in bridging the gap with supervised and stereo methods. However, these approaches focus their interest on dense 3D reconstruction and sometimes on tiny details that are superfluous for autonomous navigation. In this paper, we propose to address this issue by estimating the navigation map under a quad tree representation. The objective is to create an adaptive depth map prediction that only extract details that are essential for the obstacle avoidance. Other 3D space which leaves large room for navigation will be provided with approximate distance. Experiment on KITTI dataset shows that our method can significantly reduce the number of output information without major loss of accuracy.

Few-shot segmentation presents a significant challenge for semantic scene understanding under limited supervision. Namely, this task targets at generalizing the segmentation ability of the model to new categories given a few samples. In order to obtain complete scene information, we extend the RGB-centric methods to take advantage of complementary depth information. In this paper, we propose a two-stream deep neural network based on metric learning. Our method, known as RDNet, learns class-specific prototype representations within RGB and depth embedding spaces, respectively. The learned prototypes provide effective semantic guidance on the corresponding RGB and depth query image, leading to more accurate performance. Moreover, we build a novel outdoor scene dataset, known as Cityscapes-3i, using labeled RGB images and depth images from the Cityscapes dataset. We also perform ablation studies to explore the effective use of depth information in few-shot segmentation tasks. Experiments on Cityscapes-3ishow that our method achieves excellent results with visual and complementary geometric cues from only a few labeled examples.

Deep learning-based image understanding techniques require a large number of labeled images for training. Few-shot semantic segmentation, on the contrary, aims at generalizing the segmentation ability of the model to new categories given only a few labeled samples. To tackle this problem, we propose a novel prototypical network (MAPnet) with multiscale feature attention. To fully exploit the representative features of target classes, we firstly extract rich contextual information of labeled support images via a multiscale feature enhancement module. The learned prototypes from support features provide further semantic guidance on the query image. Then we adaptively integrate multiple similarity-guided probability maps by attention mechanism, yielding an optimal pixel-wise prediction. Furthermore, the proposed method was validated on the PASCAL-5idataset in terms of 1-way N-shot evaluation. We also test the model with weak annotations, including scribble and bounding box annotations. Both the qualitative and quantitative results demonstrate the advantages of our approach over other state-of-the-art methods.

This paper deals with new augmentation methods for an unconventional imaging modality sensitive to the physics of the observed scene called polarimetry. In nature, polarized light is obtained by reflection or scattering. Robotics applications in urban environments are subject to many obstacles that can be specular and therefore provide polarized light. These areas are prone to segmentation errors using standard modalities but could be solved using information carried by the polarized light. Deep Convolutional Neural Networks (DCNNs) have shown excellent segmentation results, but require a significant amount of data to achieve best performances. The lack of data is usually overcomed by using augmentation methods. However, unlike RGB images, polarization images are not only scalar (intensity) images and standard augmentation techniques cannot be applied straightforwardly. We propose enhancing deep learning models through a regularized augmentation procedure applied to polarimetric data in order to characterize scenes more effectively under challenging conditions. We subsequently observe an average of 18.1 % improvement in IoU between not augmented and regularized training procedures on real world data.

Monocular depth estimation is a recurring subject in the field of computer vision. Its ability to describe scenes via a depth map while reducing the constraints related to the formulation of perspective geometry tends to favor its use. However, despite the constant improvement of algorithms, most methods exploit only colorimetric information. Consequently, robustness to events to which the modality is not sensitive to, like specularity or transparency, is neglected. In response to this phenomenon, we propose using polarimetry as an input for a self-supervised monodepth network. Therefore, we propose exploiting polarization cues to encourage accurate reconstruction of scenes. Furthermore, we include a term of polarimetric regularization to state-of-the-art method to take specific advantage of the data. Our method is evaluated both qualitatively and quantitatively demonstrating that the contribution of this new information as well as an enhanced loss function improves depth estimation results, especially for specular areas.

Many fields of application could benefit from an accurate registration of measurements of different modalities over a known 3D model. However, aligning a 2D image to a 3D model is a challenging task and is even more complex when the two have a different modality. Most of the 2D/3D registration methods are based on either geometric or dense visual features. Both have their own advantages and their own drawbacks. We propose, in this paper, to mutually exploit the advantages of one feature type to reduce the drawbacks of the other one. For this, an hybrid registration framework has been designed to mutually align geometrical and dense visual features in order to obtain an accurate final 2D/3D alignment. We evaluate and compare the proposed registration method on real data acquired by a robot equipped with several visual sensors. The results highlights the robustness of the method and its ability to produce wide convergence domain and a high registration accuracy.

In the robotic field, navigation and path planning applications benefit from a wide range of visual systems (e.g, perspective cameras, depth cameras, catadioptric cameras, etc.). In outdoor conditions, these systems capture information in which sky regions cover a major segment of the images acquired. However, sky regions are discarded and are not considered as visual cue in vision applications. In this paper, we propose to estimate attitude of Unmanned Aerial Vehicle (UAV) from sky information using a polarimetric camera. Theoretically, we provide a framework estimating the attitude from the skylight polarized patterns. We showcase this formulation on both simulated and real-word data sets which proved the benefit of using polarimetric sensors along with other visual sensors in robotic applications.

This paper aims to optimize the coverage of a given area from a set of views to allow a complete mosaicing. Among the investigated methods to find the best camera positions, two of them are studied, namely the Particle Swarm Optimization (PSO) and the Genetic Algorithms (GA). After having performed experiments to compare the algorithms, the hybridization of GA and PSO is investigated. To validate the proposed method, it is simulated area of irregular shapes with the cameras mounted on a Unmanned Aerial Vehicles (UAVs). V-REP is used to simulate the UAVs in an indoor environment and satellite images are used for a large outdoor area. The simulation validates the efficiency of the proposed method to find the optimal position of cameras. Then by using the images acquired it is possible to monitor the area and to compute a full mosaic of it.

A non parametric method that defines a pixel neighborhood within catadioptric images is presented in this paper. It is based on an accurate modeling of the mirror shape by using polarization imaging. Unlike the most of current processing methods in the literature, this method is non-parametric and can deal with the deformation of catadioptric images. This paper demonstrates how an appropriate neighborhood can be derived from the polarization parameters by estimation of the degree of polarization and the angle of polarization which in return directly provide an adapted neighborhood of each pixel that can be used to perform image derivation, edge detection, interest point detection and namely image matching, all in the catadioptric image plane.

Images produced with catadioptric sensors contain a significant amount of radial distortions and variation in inherent scale. Direct application of conventional methods of image processing without considering the mirror anamorphosis yields erroneous issues. Examples of such applications are those based on the pixel neighborhood: one can mention image derivation, image convolution or image matching. The inefficiency of classical algorithms in catadioptric images processing leads in general to biased results. In this paper, we propose a new method to match a pair of catadioptric images taking into account their distorted topology. This method is based on an adapted window resulting from a convenient parameterization of the mirror. Experiments have been performed on image sequences acquired with a parabolic catadioptric sensor.

A new efficient method of calibration for catadioptric sensors is presented in this paper. It is based on an accurate reconstruction of the mirror by means of polarization imaging and it permits to relax most of the constraints related to the calibration of the catadioptric systems. No image processing and no calibration pattern are required so that the calibration can be performed in "one click". As a result, the coordinates of the projection rays and a mapping between projection rays and image pixels is obtained. Besides, two methods of triangulation have been adapted to catadioptric sensors and tested. Experiments prove the validity of the method with some preliminary results on 3D reconstruction.

We present a new way of solving the ambiguity that appears in the shape from polarization method. This ambiguity, concerning the normals orientation of the surface, was previously solved thanks to a propagation algorithm. In this paper, the ambiguity is solved thanks to a special active lighting. This method is less time computing and is more robust to noise acquisition. We first demonstrate how to compute the right normals on the whole surface. Then, we show an application on the 3D reconstruction of a specular metallic object made by stamping and polishing.

This paper describes an experiment of recognition of sequences of digits pronounced continuously. This experiment takes place in the design of a practical system for postal codes recognition. The speech signal is analyzed by a 15-channel spectral analyzer. The digit sequence is then recognized using a semi-global procedure which consists of two successive steps: - segmentation of the sequence into the digits which comprise the string, - recognition of the individual digits by a dynamic programming technique which was previously designed in the laboratory. Typical Results of segmentation and recognition will be given and discussed.