François Jouen

Appearance-based gaze tracking with spectral clustering and semi-supervised Gaussian process regression

Two of the challenges in appearance-based gaze tracking are: 1) prediction accuracy, 2) the efficiency of calibration process, which can be considered as the collection and analysis phase of labelled and unlabelled eye data. In this paper, we introduce an appearance-based gaze tracking model with a rapid calibration. First we propose to concatenate local eye appearance Center-Symmetric Local Binary Pattern(CS-LBP) descriptor for each subregion of eye image to form an eye appearance feature vector. The spectral clustering is then introduced to get the supervision information of eye manifolds on-line. Finally, taking advantage of eye manifold structure, a sparse semi-supervised Gaussian Process Regression(GPR) method is applied to estimate the subjects gaze coordinates. Experimental results demonstrate that our system with an efficient and accurate 5-points calibration not only can reduce the run-time cost but also can lead to a better accuracy result of 0.9°.

Kinaesthetic and visual perceptions of orientations.

In the present study we compare the kinaesthetic and visual perception of the vertical and horizontal orientations (subjective vertical and subjective horizontal) to determine whether the perception of cardinal orientations is amodal or modality-specific. The influence of methodological factors on the accuracy of perception is also investigated by varying the stimulus position as a function of its initial tilt (clockwise or counterclockwise) and its angle (22°, 45°, 67°, and 90°) in respect to its physical orientation. Ten participants estimated the vertical and horizontal orientations by repositioning a rod in the kinaesthetic condition or two luminous points, forming a ‘virtual line’ in the visual condition. Results within the visual modality replicated previous findings by showing that estimation of the physical orientations is very accurate regardless of the initial position of the virtual line. In contrast, the perception of orientation with the kinaesthetic modality was less accurate and systematically influenced by the angle between the initial position of the rod and the required orientation. The findings question the assumption that the subjective vertical is derived from an internal representation of gravity and highlight the necessity of taking into account methodological factors in studies on subjective orientations.

Kinesthetic Perception of the Vertical as a Function of Sport Experience.

Accurate perception of the vertical is critical to spatial orientation, which essential to ensure that behavior isappropriately adapted to the environment. As such, it is surprising that our understanding ofvertical perception and the frames ofreference used to orient oneself in space is less than adequate. This paper attempts to redress this situation by showing a link between experiences influencing posture and the perception ofverticality. This link has implications for understanding how humans perceive the spatial layout of their environment and their relation to that layout. To define an orientation in space, different frames of reference can be selected, with a broad distinction between egocentric (referenced to the participants body) and allocentric (referenced to environmental cues) spatial frames ofreference (for a review,see Wade, 1992). The physical vertical (i.e., the line through the mass centroid of the earth) is considered a primary allocentric reference, whereas the posturalvertical (i.e., the orienta-