Multi-depth cross-calibration of remote eye gaze trackers and stereoscopic scene systems

Abstract

Abstract-We present a robust and accurate technique for the cross-calibration of 3D remote gaze trackers with stereoscopic scene vision systems between which no common imaging area exists. We empirically demonstrate that a multidepth calibration approach yields remarkably superior results for obtaining 3D Point-of-Gaze (PoG) when compared with traditional methods using monocular scene cameras and coplanar eye gaze calibration points.

Predicting driver behaviour at intersections based on driver gaze and traffic light recognition

“…The method for obtaining the distance between the TL and the driver is determined from the depth data in the images and from the stereo camera calibration data. From the stereo cameras used to collect the driving sequence data, we have their calibration matrix [30], along with the baseline and focal length information of the stereo cameras. The distance can then be computed using the methods outlined in [28, 30].…”

“…The feature vector for an image then consists of four elements: point of gaze, depth of gaze, head position, state of TL. The gaze data of the driver is extracted using the method used by Kowsari et al [30]. The eye tracker data is transformed to be interpreted as a 3D line of gaze vector in the forward stereo camera coordinate system.…”

A Review and Analysis of Eye-Gaze Estimation Systems, Algorithms and Performance Evaluation Methods in Consumer Platforms

“…A portable driver assistance system involving driver drowsiness detection and eye gaze tracking is implemented using a Raspberry Pi and machine vision algorithms in [130]. A new multi-depth calibration approach is presented in [131]for obtaining 3D user PoG with stereo face cameras and monocular scene cameras for performing driver intent and actions prediction. Several works study the dynamics between head pose and gaze behavior of drivers.…”

Detection and recognition of traffic signs inside the attentional visual field of drivers

“…In this work, our main focus lies on detection and recognition of signs within the visual field of the driver. In order to relate the 3D Line-of-Gaze (LoG) of the driver to the depth map obtained by the forward stereo camera system and derive the 3D Point-of-Gaze (PoG), we used a technique proposed in our laboratory [84] to identify the 3D PoG in absolute coordinates expressed in the frame of reference of the vehicle. Figure 4.1 depicts the remote eye tracking system and the stereoscopic vision system.…”

“…The eye tracking system and the forward stereoscopic system each operate in their own frame of reference, and we need to transform the 3D circle corresponding to the visual attentional area of the driver into the frame of reference of the stereo system of the vehicle. To accomplish this, we used a cross-calibration process developed earlier in our laboratory to compute the parameters of the rigid transformation (a translation T and a rotation matrix R) between the eye tracker and the stereoscopic system [84]:…”