Eye Location Research Articles

Intelligent Regulation Algorithm of Automobile Rear View Mirror based on Eye Location

Intelligent Regulation Algorithm of Automobile Rear View Mirror based on Eye Location

Inner-Canthus Localization of Thermal Images in Face-View Invariant

Inner-canthus localization has played an essential role in measuring human body temperature. This is due to the theory that human core body temperature can be measured in the inner-canthus. Such measurement is useful for mass screening since it is non-contact, non-invasive and fast. This paper presents an algorithm that has been developed to locate the inner-canthus. The algorithm proposed a robust method in various face-view, i.e., frontal, sided and tilted. The algorithm consisted of: face segmentation, determining face-orientation, rotating face into straight view, eye localization, and inner-canthus localization. The face segmentation used human temperature threshold of 34°C — the face orientation used trend line of a middle point between each most-bottom and most-top coordinates. The face rotation was based on the gradient of the trend line. Once the face is rotated, the eye location was determined using facial proportion. The inner-canthus location was determined as the highest intensities in the eye-frame. The test on 15 thermal images of faces with various view showed localization accuracy of 80% for eye-frame determination and 100% for inner-canthus localization.

Inner-Canthus Localization of Thermal Images in Face-View Invariant

Inner-canthus localization has played an essential role in measuring human body temperature. This is due to the theory that human core body temperature can be measured in the inner-canthus. Such measurement is useful for mass screening since it is non-contact, non-invasive and fast. This paper presents an algorithm that has been developed to locate the inner-canthus. The algorithm proposed a robust method in various face-view, i.e., frontal, sided and tilted. The algorithm consisted of: face segmentation, determining face-orientation, rotating face into straight view, eye localization, and inner-canthus localization. The face segmentation used human temperature threshold of 34°C — the face orientation used trend line of a middle point between each most-bottom and most-top coordinates. The face rotation was based on the gradient of the trend line. Once the face is rotated, the eye location was determined using facial proportion. The inner-canthus location was determined as the highest intensities in the eye-frame. The test on 15 thermal images of faces with various view showed localization accuracy of 80% for eye-frame determination and 100% for inner-canthus localization.

The cranio-mandibular complex of the nightjar Systellura longirostris (Aves, Caprimulgiformes): functional relationship between osteology, myology and feeding

Here we address the morphology of the cranial musculoskeletal in the South American Band-winged nightjar (Systellura longirostris, Caprimulgiformes, Aves) with the focus upon in how the mouth closure and opening mechanism is. Bony characters such as the thinness of the interorbital septum and the reduction of the postorbital process co-assist the location of large eyes, although correlations between eye size brain mass and body mass revealed that the Band-winged nightjar’s eye is bigger than expected. The reduction of the processus orbitalis of the os quadratum would allow the great mouth opening given by a large maxillary protraction. Systellura longirostris shows a zona flexoria intramandibularis with marked reduction in bone ossification that helps to the lateral opening of the jaws. In Systellura longirostris, the m. pseudotemporalis profundus, m. adductor mandibulae externus superficialis, and m. adductor mandibulae externus medialis are absent. The adductor muscles represent 66.614% of the total mandibular muscle mass. Physiological Cross Sectional Area (PCSA) values of the adductor muscles are high and the fibers are relatively long compared to other zoophagous birds, another condition that favors adduction. The m. adductor mandibulae externus profundus is the largest jaw muscle regarding its mass but its osteological correlate on the neurocranium, the fossa temporalis, is absent. In the absence of some muscles in charge of the bill closing, an elastic instability model for jaw closing is proposed.

Quiet Eye: Practical Applications in Sport and Physical Education

Visual behaviors are often overlooked in sport performance training. The literature suggests that training visual behaviors, such as the quiet eye, can result in better performance across a wide variety of motor skills. The purpose of this article is to provide specific recommendations to physical education teachers and coaches for including quiet-eye training in their practices. Studies have shown that longer quiet-eye durations are associated with improved sport performance, but the location of the quiet eye varies from sport to sport. Studies also indicate that experts, novices and children benefited from quiet-eye training, and the effects of the training applied to experimental, competitive and heightened-anxiety situations involving many sports and skills. It is recommended that physical education teachers and coaches include quiet-eye training for the acquisition and improvement of motor skills.

Location of Eyes in Images of Human Faces through Analysis Variance Shine Intensity

Extraction of facial features is an important step in automatic visual interpretation and recognition of human faces. Among the facial features, the eyes play a major role in the recognition process. In this article, we present an approach to detect and locate the eyes in frontal face images. Eye regions are identified using the technique of voucher detection based on mathematical morphology. After this identification is made a comparison between the variances of three different portions of each candidate region to eye (set of pixels belonging to the candidate region as a whole, set of pixels contained in a minimum rectangle circumscribed to the candidate region and set of pixels of the candidate region belonging to a horizontal band that crosses the center of mass of this region). The calculation of these variances also considers the R, G, and B channels, as well as the gray version of the input image.

A Global Climatology of Tropical Cyclone Eyes

Abstract Intense tropical cyclones (TCs) generally produce a cloud-free center with calm winds, called the eye. The Automated Rotational Center Hurricane Eye Retrieval (ARCHER) algorithm is used to analyze Hurricane Satellite (HURSAT) B1 infrared satellite imagery data for storms occurring globally from 1982 to 2015. HURSAT B1 data provide 3-hourly observations of TCs. The result is a 34-yr climatology of eye location and size. During that time period, eyes are identified in about 13% of all infrared images and slightly more than half of all storms produced an eye. Those that produce an eye have (on average) 30 h of eye scenes. Hurricane Ioke (1992) had the most eye images (98, which is 12 complete days with an eye). The median wind speed of a system with an eye is 97 kt (50 m s−1) [cf. 35 kt (18 m s−1) for those without an eye]. Eyes are much more frequent in the Northern Hemisphere (particularly in the western Pacific) but eyes are larger in the Southern Hemisphere. The regions where eyes occur are expanding poleward, thus expanding the area at risk of TC-related damage. Also, eye scene occurrence can provide an objective measure of TC activity in place of those based on maximum wind speeds, which can be affected by available observations and forecast agency practices.

Eye tracker accuracy: quantitative evaluation of the invisible eye center location.

We present a new method to evaluate the accuracy of an eye tracker-based eye localization system. Measuring the accuracy of an eye tracker's primary intention, the estimated point of gaze, is usually done with volunteers and a set of fixation points used as ground truth. However, verifying the accuracy of the location estimate of a volunteer's eye center in 3D space is not easily possible. This is because the eye center, the center of corneal curvature, is an intangible point. We evaluate the eye location accuracy by using an eye phantom instead of eyes of volunteers. For this, we developed a testing stage with a realistic artificial eye and a corresponding kinematic model, which we trained with [Formula: see text] data. This enables us to precisely evaluate the eye location estimate of an eye tracker. We show that the proposed testing stage with the corresponding kinematic model is suitable for such a validation. Further, we evaluate a particular eye tracker-based navigation system and show that this system is able to successfully determine the eye center with a mean accuracy of 0.68 mm. We show the suitability of the evaluated eye tracker for eye interventions, using the proposed testing stage and the corresponding kinematic model. The results further enable specific enhancements of the navigation system to potentially get even better results.

Imagining the Unimaginable Faces by Deconvolutional Networks.

We tackle the challenge of constructing 64 pixels for each individual pixel of a thumbnail face image. We show that such an aggressive super-resolution objective can be attained by taking advantage of the global context and making the best use of the prior information portrayed by the image class. Our input image is so small (e.g., pixels) that it can be considered as a patch of itself. Thus, conventional patch-matching-based super-resolution solutions are unsuitable. In order to enhance the resolution while enforcing the global context, we incorporate a pixel-wise appearance similarity objective into a deconvolutional neural network, which allows efficient learning of mappings between low-resolution input images and their high-resolution counterparts in the training data set. Furthermore, the deconvolutional network blends the learned high-resolution constituent parts in an authentic manner, where the face structure is naturally imposed and the global context is preserved. To account for the possible artifacts in upsampled feature maps, we employ a sub-network composed of additional convolutional layers. During training, we use roughly aligned images (only eye locations), yet demonstrate that our network has the capacity to super-resolve face images regardless of pose and facial expression variations. This significantly reduces the requirement of precisely face alignments in the data set. Owing to the network topology we apply, our method is robust to translational misalignments. In addition, our method is able to upsample rotational unaligned faces with data augmentation. Our extensive experimental analysis manifests that our method achieves more appealing and superior results than the state of the art.

30‐1: Cyclopean Eye based Binocular Orientation in Virtual Reality

Currently the individual variability in the Cyclopean Eye's position is not considered when presenting virtual reality content. We proposed a novel perceptual space model for virtual reality content based on the location of the Cyclopean Eye. The proposed methodology may improve the immersive viewing experience and interaction accuracy between virtual reality system and individual user.

Noninvasive eye localization in ocular proton therapy through optical eye tracking: A proof of concept.

Over the last four decades, Ocular Proton Therapy has been successfully used to treat patients affected by intraocular lesions. For this, treatment geometry verification is routinely performed using radiographic images to align a configuration of fiducial radiopaque markers implanted on the sclera outer surface. This paper describes the clinical application of an alternative approach based on optical eye tracking for three-dimensional noninvasive and automatic eye localization. An experimental protocol was designed to validate the optical-based eye referencing against both radiographic imaging system and the clinically used EYEPLAN treatment planning system. The eye tracking system (ETS) was installed in the OPTIS 2 treatment room at PSI to acquire eye motions during the treatment of nine patients. The pupil position and the cornea curvature center were localized by segmenting the pupil contour and corneal light reflections on the images acquired by a pair of calibrated optical cameras. After calibration of the ETS, a direct comparison of radiopaque markers position, and consequentially eye position and orientation, provided by the ETS, radiographs and EYEPLAN was performed. Nineteen out of thirty total monitored fractions were excluded from the study due to poor visibility of corneal reflection, resulting in a success rate of acquisition of 37%. For these data, overall median agreement between ETS-based and x-ray-based markers position assessment were 0.29mm and 0.94° for translations and rotations, respectively. Small discrepancies were also measured in the eye center estimates of the ETS and EYEPLAN. Conversely, variations in measured eye orientation were higher, with interquartile range (IQR) between 4.39° and 7.58°. Nonetheless, dosimetric evaluation of the consequence of ETS uncertainties showed that the target volume would still be covered by more than 95% of the dose in all cases. An ETS was successfully installed in a clinical ocular proton therapy treatment room and used to monitor eye position and orientation in a clinical scenario. First results show the potential of such a system as an eye localization device. However, the low success rate prevents straightforward clinical application and needs further improvements aimed at increasing corneal reflection visibility.

Fully-automated alignment of 3D fetal brain ultrasound to a canonical reference space using multi-task learning.

Methods for aligning 3D fetal neurosonography images must be robust to (i) intensity variations, (ii) anatomical and age-specific differences within the fetal population, and (iii) the variations in fetal position. To this end, we propose a multi-task fully convolutional neural network (FCN) architecture to address the problem of 3D fetal brain localization, structural segmentation, and alignment to a referential coordinate system. Instead of treating these tasks as independent problems, we optimize the network by simultaneously learning features shared within the input data pertaining to the correlated tasks, and later branching out into task-specific output streams. Brain alignment is achieved by defining a parametric coordinate system based on skull boundaries, location of the eye sockets, and head pose, as predicted from intracranial structures. This information is used to estimate an affine transformation to align a volumetric image to the skull-based coordinate system. Co-alignment of 140 fetal ultrasound volumes (age range: 26.0 ± 4.4 weeks) was achieved with high brain overlap and low eye localization error, regardless of gestational age or head size. The automatically co-aligned volumes show good structural correspondence between fetal anatomies.

Eye landmarks detection via two-level cascaded CNNs with multi-task learning

Despite extensive researches for eye localization, accuracy of eye landmarks detection is easily influenced by illumination and the difference in eye state conversion between open and close. To address these problems, we present a novel approach for eye landmarks detection with two-level cascaded convolutional neural networks. Network at the first level utilizes eye state estimation as auxiliary task to provide initial positions of the eye. The shallower network at the second level fine tunes eye positions by taking small regions centered at predicted eye points as input. To train our model, we introduce OCE dataset, the first dataset with the eye in different states. Our method achieves mean detection error of 5.6% on OCE dataset. Further experiments are tested on UBIRIS V1.0, MMU V1.0 and MICHE-I and their results demonstrate acceptable performance of our method.

Detecting Driver Drowsiness Using Eye State

Various studies have suggested that 20%-50% road accidents are fatigue-related. Nevertheless, on certain roads, the maximum value is out of scope. This paper aims at developing a vision-based drowsiness detecting system which is noninvasive and accurate to help alert drowsy driver in advance. Eye state is used to evaluate drowsiness. First, face region is detected through AdaBoost with Haar feature. Then instead of applying old fashioned eye location methods which many drowsiness detect systems still possess, our system obtain eye region through a customized state of art face alignment algorithm based on Regressing Local Feature. Finally, an SVM is trained to distinguish eye states. Results demonstrate that our system can achieve a high detection rate in various lighting circumstances.

Research on eye localisation method based on adaptive correlation filter

Eye location is one of the most important steps in the face recognition and visual tracking system. This article combines the correlation filter with integral projection to detect eye position precisely, and puts forward two improvements in the training and test phases. Firstly, the adaptive synthetic correlation filter are rotated within the angle range of –0.2, –0.1, 0, 0.1, 0.2, and the corresponding location of the maximum grey value was served as an initial anchor point. Secondly, within the anchor point 5 × 5 neighbourhood, grey-level integral projection is implemented in the horizontal and vertical direction respectively. Finally, the ultimate goal orientation chooses the position of minimum integration. Compared with the average synthetic exact filter, minimum output sum of squared error filter and adaptive synthetic correlation filter, the experimental results show the proposed improvements are superior to the general algorithm in the accuracy and stability.

Research on eye localization method based on adaptive correlation filter

Eye location is one of the most important steps in the face recognition and visual tracking system. This article combines the correlation filter with integral projection to detect eye position precisely, and puts forward two improvements in the training and test phases. Firstly, the adaptive synthetic correlation filter are rotated within the angle range of –0.2, –0.1, 0, 0.1, 0.2, and the corresponding location of the maximum grey value was served as an initial anchor point. Secondly, within the anchor point 5 × 5 neighbourhood, grey-level integral projection is implemented in the horizontal and vertical direction respectively. Finally, the ultimate goal orientation chooses the position of minimum integration. Compared with the average synthetic exact filter, minimum output sum of squared error filter and adaptive synthetic correlation filter, the experimental results show the proposed improvements are superior to the general algorithm in the accuracy and stability.

Real Time Eye Detector with Cascaded Convolutional Neural Networks

An accurate and efficient eye detector is essential for many computer vision applications. In this paper, we present an efficient method to evaluate the eye location from facial images. First, a group of candidate regions with regional extreme points is quickly proposed; then, a set of convolution neural networks (CNNs) is adopted to determine the most likely eye region and classify the region as left or right eye; finally, the center of the eye is located with other CNNs. In the experiments using GI4E, BioID, and our datasets, our method attained a detection accuracy which is comparable to existing state-of-the-art methods; meanwhile, our method was faster and adaptable to variations of the images, including external light changes, facial occlusion, and changes in image modality.

Hybrid sensing face detection and registration for low-light and unconstrained conditions.

The capability to track, detect, and identify human targets in highly cluttered scenes under extreme conditions, such as in complete darkness or on the battlefield, has been one of the primary tactical advantages in military operations. In this paper, we propose a new collaborative, multi-spectrum sensing method to achieve face detection and registration under low-light and unconstrained conditions. We design and prototype a novel type of hybrid sensor by combining a pair of near-infrared (NIR) cameras and a thermal camera (a long-wave infrared camera). We strategically surround each NIR sensor with a ring of LED IR flashes to capture the "red-eye," or more precisely, the "bright-eye" effect of the target. The "bright-eyes" are used to localize the 3D position of eyes and face. The recovered 3D information is further used to warp the thermal face imagery to a frontal-parallel pose so that additional tasks, such as face recognition, can be reliably conducted, especially with the assistance of accurate eye locations. Experiments on real face images are provided to demonstrate the merit of our method.

High-boost Weber local filter for precise eye localization under uncontrolled scenarios

The task of precise eye localization for a given face is very important for many computer vision applications such as face recognition, face and eye tracking, and face alignment. Further, this is a challenging task when performed in uncontrolled scenarios. This paper proposes a new method for precise eye localization in uncontrolled scenarios using the high-boost Weber local filter (HBWLF); the proposed filter emphasizes high-frequency components, without eliminating the low-frequency ones, and thus enhances the details of the periocular region (region around the eyes), which is critical to the task of eye localization. The proposed method was evaluated using Labeled Faces in the Wild (LFW) and BioID databases, and the results show that HBWLF yields better performance than several classical methods.

A Spiking Neural Networks Based Face Recognition Algorithm

In this paper, the complexity algorithm is used to locate the human eyes, and then the best threshold method is used to locate the human eyes accurately. This method is more accurate than the gray projection method, and it is faster and less affected by the light and noise. Spiking neural networks, which inherit the parallel mechanism from biological system, are used to extract the face features. The spiking neural networks can remember key features of a visual image through synapse strength distribution and recall the visual image by triggering a specific neuron. Based on the key features, the nearest neighbor classifier is used for matching faces. Experimental results show that the proposed algorithm of eye location works well and has advantages about eye location in multi-position and complex background, and the face features extraction based on spiking neural networks can achieve high recognition rate and reduce the time. Furthermore, the algorithm can be transformed to GPU platform and can be speed up dramatically.