Search

International audience; Monocular augmented reality devices are used in aviation to help civilian or military pilots in their flying task. In those devices, a virtual image is projected in front of one eye on a see-through glass allowing them to see environment with both eyes. This glass is mounted on a helmet allowing users to maintain the virtual information, usually find in the primary flight display, in front of the eye. Given that the image is presented in front of only one eye, the two eyes are not stimulated in the same way and it can creates a phenomenon known as binocular rivalry. When it appeared, brain is not able to merge the two visuals information and an alternation between them can occur. This alternation appears according to visual condition and the question arises as is it relevant to choose the eye to display the image to limit binocular rivalry and guarantee a good performance in the tasks of recognition and control. The control task consists in matching a monocular item with a binocular item and the detection task consists in giving the right opening orientation of Landolt rings that can be monocular or binocular. Our study aims to compare the performances as a function of the position of the virtual image. These results are then compared to the results of several dominant eye tests to determine if one test can objectively determine on which eye the monocular information should be displayed when using a monocular see-through device.

International audience

International audience

International audience

International audience; To perform tasks, such as detection or recognition of objects in natural environment by day or by night, it is possible to use images acquired from different sensors: natural images, thermal images from infrared sensor or images acquired during the night with light intensifier. Our goal is to improve the efficiency of operators performing these tasks by providing a synthetic image made from different sensors that will enhance the information content of each sensor. First we have to know the image characteristics we use: edges detection and spatial frequencies are statistically analyzed. They show the differences between the image sensors. Then we have to understand which information is important for the observer, from each sensor, for a given task. To obtain this knowledge, we have developed a psychovisual experiment to discriminate vehicles, by using the method of [Gosselin, Schyns, 2001, Vision Research 41, 2261-2271]. Stimuli presented to the observers are constructed by filtering the original image at different scales and multiplied by Gaussian “bubbles” that partially obscure the signal [Lelandais, Plantier, BIOSIGNALS2013, Spain]. The results of the psychovisual experiment give the number of bubbles necessary to perform the task and to determine the useful parts of vehicles for their discrimination.

International audience; One way in psychovisual experiment to understand human visual system is to analyze separately contents of different spatial frequency bands. To prepare images for this purpose, we proceed to a decomposition of the original image by a wavelet transform centered on selected scales. The wavelets used are Difference Of Gaussians (DOG) according to works modeling the human visual system. Before rebuilding the visual stimulus, various transformations can be performed on different scales to measure the efficiency of the observer, for a given task, according to the spatial frequencies used. The problem is that if we use an incomplete wavelet basis during decomposition, there is a significant loss of information between the original image and the reconstructed image. The work presented here offers a way to solve this problem by using coefficients appropriate for each scale during the decomposition step.

National audience

National audience; One challenging human factors issue in Helmet-Mounted Displays is the alteration of perceived distance. Distances judgments in virtual environments and Helmet-Mounted Displays are generally underestimated compared to the real world. Indeed, visual stimulation in Helmet-Mounted Displays presents some discrepancy with natural visual stimulation. After a brief review of the literature pertaining to the representation of space in Helmet-Mounted Displays, we explore two potential factors for the altered distance perception: the increased interocular separation (also called hyperstereopsis) and the reduction of the field of view. Some laboratory and training ground data are reported. This research has direct impact on the design and use of Helmet-Mounted Displays in order to ensure flight safety.

International audience

Les images de notre quotidien du 21ème siècle sont acquises et visionnées en couleur. Nos télévisions, nos téléphones et ordinateurs, tous permettent une vision indirecte, de la scène à l’œil en passant par un système de d’acquisition et une système d’affichage, avec l’illusion d’une concordance optimale avec la vision directe (de la scène à l’œil) que nous portons au monde qui nous entoure. Pour certaines applications d’imagerie extrêmement exigeantes en termes de colorimétrie, telles que : l’industrie du cinéma, la surveillance, l’industrie du textile ou bien encore l’exploration spatiale, cette illusion est régulièrement altérée par les conditions extrêmes d’acquisition rencontrées. L’objectif de cette thèse est de définir des méthodes pour la reproduction d’images numériques de très haute qualité colorimétrique. Deux axes de recherche principaux sont considérés: la sensibilité spectrale de l’œil humain, et la chaîne du traitement de l’information couleur par les capteurs d’images modernes. Les deux premiers chapitres du manuscrit donnent les connaissances et méthodes liées à ces deux axes. Dans ce cadre, deux études ont été menées sur (i) la méthode de calcul de la matrice de correction des couleurs, et (ii) la prise en compte des variations individuelles de la vision des couleurs par l’humain. Ce travail de thèse vise à définir si l’un de ces deux aspects de la chaîne de la vision indirecte, souvent étudiés indépendamment l’un de l’autre, est à l’origine de trouble importants pour la fidélité colorimétrique. Pour répondre à cette problématique, deux études ont été réalisées: la première quantifie les erreurs résiduelles après l’application à une image de corrections calculées selon deux méthodes calculatoires (moindres carrés linéaires et régression polynomiale d’ordre 2). L’étude se penche en particulier sur deux types de scènes: celles qui présentent des particularités spectrales restreintes, et celles ne permettant pas l’utilisation de mires de calibration. Les résultats sont issus de l’étude de deux capteurs d’images CMOS et sous illuminant D65. Ils indiquent en bon accord avec les études précédentes que pour des scènes contenant des couleurs de l’ensemble du domaine spectral visible, le nombre de patches utilisés pour le calcul de la matrice de correction des couleurs (CCM) peut être réduit en dessous de 24 (mire ColorChecker® X-rite). Concernant les scènes présentant des particularités spectrales, des corrections teintes-spécifiques ont été appliquées à partir de la mire Next Generation Target (NGT). Les résultats suggèrent de compléter l’étude par d’autres méthodes de sélection des teintes pour le calcul de CCM. Dans les applications qui ne permettent pas l’utilisation de mires de calibration, une méthode de caractérisation par imageur multispectral 6-bandes (déjà disponible sur de nombreuses applications spatiales) est étudiée. Elle présente des résultats acceptables sur la caractérisation de la mire ColorChecker® X-rite qui présente des réflectances qui varient de manière très progressive sur le domaine du visible. La deuxième étude quantifie les différences qui peuvent être engendrées par les variations individuelles de la vision humaine, dernier élément de la chaine de la vision indirecte. Les résultats issus de données en accès-libre dans la littérature, montrent que pour certaines teintes les différences de vision des couleurs peuvent s’avérer très critiques et suggèrent le développement de méthodes d’individualisation des images. Une revue est ensuite faite concernant les leviers technologiques utilisés pour le développement d’une expérience d’égalisation visant à mesurer la sensibilité spectrale d’individus soumis à ce contexte d’exigences colorimétriques importantes. Une discussion est alors ouverte concernant le développement de nouvelles expériences d’égalisations, proposant des alternatives pratiques issues des travaux de la littérature très récente. The images of our daily life are acquired and viewed in color. Our televisions, smartphones and computers, all allow for indirect vision: from the scene to the eye, through an acquisition system and a display system. This is done with the illusion of an optimal match with the direct vision (from the scene to the eye) that we have of our surrounding world. For color-critical imaging applications, such as the cinema industry, surveillance, the textile industry or space exploration, this illusion is regularly altered by the extreme acquisition conditions encountered.The objective of this thesis is to define methods for the reproduction of digital images of very high color quality. Two main research axes are considered: the spectral sensitivity of the human eye, and the processing chain of color information by modern image sensors. The first two chapters of the manuscript give the knowledge and methods related to these two axes. In this context, two studies were conducted on (i) the method of calculating the color correction matrix, and (ii) the consideration of individual variations of color vision by humans. This work aims at defining if one of these two aspects of the chain of the indirect vision, often studied independently of each other, is at the origin of important disorders for the colorimetric fidelity.To address this issue, two studies have been conducted: the first quantifies the residual errors after the application to an image of corrections calculated by two computational methods (linear least squares and polynomial regression of order 2). The study focuses on two types of scenes: those with restricted spectral features, and those that do not allow the use of calibration charts. The results are based on the study of two CMOS image sensors and under illuminant D65. They indicate in good agreement with the previous studies that for scenes containing colors of the whole visible spectral range, the number of patches used for the calculation of the color correction matrix (CCM) can be reduced below 24 (ColorChecker® X-rite chart). On the other hand a reduction in the number of patches does not allow the use of the polynomial regression method admitted by the literature as more effective on complete scenes. For scenes with spectral peculiarities, hue-specific corrections were calculated and applied from the Next Generation Target (NGT) test pattern. The results suggest to complete the study with other methods of selecting the selected hues for the calculation of CCM. Other methods for calculating CCM should also be analyzed. In applications that do not allow the use of calibration charts, a method for scene characterization with a multispectral 6-band imager (already available on many space applications) is studied. It shows acceptable results on the characterization of the ColorChecker® X-rite test chart which presents reflectances that vary in a very progressive way in the visible range.The second study quantifies the differences which can be generated by the individual variations of the human vision, last element of the chain of the indirect vision. Results from open access data in the literature show that for some hues the differences in color vision can be very critical and suggest the development of methods for image individualization. A review is then made concerning the technological levers used for the development of an equalization experiment aiming at measuring the color matching functions of individuals in the context of color-critical applications. A discussion is then opened concerning the development of new equalization experiments, proposing practical alternatives resulting from the works of the very recent literature.