70 JOURNAL OF DISPLAY TECHNOLOGY, VOL. 8, NO. 2, FEBRUARY 2012

Accommodative Response of Integral Imagingin Near Distance

Youngmin Kim, Jongshin Kim, Keehoon Hong, Hee Kyung Yang, Jae-Hyun Jung, Heejin Choi, Sung-Wook Min,Jong-Mo Seo, Jeong-Min Hwang, and Byoungho Lee

Abstract—Objective evaluation results using optometric deviceto measure accommodative responses in viewing a real objectand integral imaging are presented. From the empirical resultsbetween the real object and an integrated three-dimensional (3D)image, we find that over 73% of participants keep eyes on the realintegrated 3D image instead of a display panel. The results alsoshow the participants do not recognize a mismatch between the ac-commodative response and the convergence of the eye, which usedto be believed as one of the major factors to cause visual fatiguein viewing near-distance integral imaging. Seventy-one normaladult subjects (23 38 years old) participated in the experiment,and accommodative response measurement results of the realintegrated image show a statistically significant concordance withreal objects.

Index Terms—Display human factors, geometrical optics, three-dimensional (3D) displays.

I. INTRODUCTION

T HE three-dimensional (3D) display technique is one ofthe most promising ways to express real images. Unlike

the other next-generation display techniques such as an organiclight-emitting diode (OLED), flexible display, and ultrathindisplay, only the 3D display technique can depict an exquisiteimage of natural world [1]–[4]. Naturalness of the reproduced3D image varies with the 3D display methods. In an autostereo-scopic display method based on either a lenticular lens or aparallax barrier, there is a tradeoff between the number of viewsand resolution of the reconstructed 3D image. Therefore, thelimitation of the number of views makes the autostereoscopicdisplay unnatural, while it is easy to fabricate the lenticular lensor parallax barrier. Holographic display, which is known as oneof the most progressive 3D display methods, can reconstruct a

Manuscript received March 15, 2011; revised July 20, 2011; accepted July 26,2011. Date of current version January 25, 2012. This work was supported by theBrain Korea 21 Program (Information Technology of Seoul National University)of the Ministry of Education, Science and Technology of Korea.

Y. Kim, K. Hong, J.-H. Jung, J.-M. Seo, and B. Lee are with the Schoolof Electrical Engineering, Seoul National University, Seoul 151-744, Korea(e-mail: rainmaker_00@naver.com; khhong@snu.ac.kr; neoengine@snu.ac.kr;callme@snu.ac.kr; byoungho@snu.ac.kr).

J. Kim is with the Graduate School of Medical Science and Engineering,KAIST, Daejeon 305-701, Korea (e-mail: jongshin.kim@gmail.com).

H. K. Yang, and J.-M. Hwang are with Department of Ophthalmology,Seoul National University College of Medicine, Seoul National UniversityBundang Hospital, Gyeonggi-do 463-707, Korea (e-mail: nan282@snu.ac.kr;hjm@snu.ac.kr).

H. Choi is with Department of Physics, Sejong University, Seoul 143-747,Korea (e-mail: hjchoi@sejong.ac.kr).

S.-W. Min is with Department of Information Display, Kyung Hee University,Seoul 130-701, Korea (e-mail: mins@khu.ac.kr).

Color versions of one or more of the figures are available online at http://ieeexplore.ieee.org.

Digital Object Identifier 10.1109/JDT.2011.2163701

3D image so that it appears as if the regenerated object is in thereal world. However, the coherent light source is necessary ingeneral, and it is still hard to reproduce full-color holography[5]. Integral imaging is a potential alternative of stereoscopyand holography. It uses an array of small lenses that are spher-ical, square, or hexagonal to produce the 3D images which canprovide both horizontal and vertical parallax, resulting from atwo-dimensional (2D) lens array [3], [6]. It provides differentperspectives according to view directions although the spatialand angular resolutions are in a tradeoff relation.

With the progress of the 3D displays, visual fatigue related tothe 3D display has recently emerged again as an important issue.Many stereoscopic films were produced in the early 1950s; how-ever, the public interests rapidly declined. One of the main rea-sons was associated with the visual fatigue caused by poor 3Deffects. Since then, great attention has been paid to clarify thecause of visual fatigue related to the 3D images. Researches con-centrated on the 3D images, especially stereoscopic displays,revealed that causes of visual fatigue included anomalies ofbinocular vision, distortion error of the eyes, conflict betweenaccommodation and convergence of the eye, and an excessivebinocular parallax. Among them, conflict between accommo-dation and convergence of the eye has been mentioned as oneof the controversial issues because of the unnatural combina-tion of stimuli. The conflict occurs because the accommodativeresponse of the eye remains fixed on the display panel, and theconvergence response varies in depth depending on the disparitydegrees of display panel.

Numerous research has attempted to expound an eye fatiguemechanism caused by mismatch of accommodation andconvergence in the stereoscopic display [7]–[13]. Despitemany attempts to explain the mechanism of eye fatigue instereoscopic displays, little research has been conducted bothsubjectively and objectively in the field of integral imaging.Recently, it was confirmed that there was no significant visualfatigue difference between integral imaging method and con-ventional binocular method by ten observers [14]. It was anobjective evaluation based on biological reactions, such as thecerebral blood flow observation, the autonomic nervous system,and an endocrine system [14]. These methods were based onindirect measurement of the eye fatigue which was obtainedfrom a biological reaction instead of the direct measurement.It is almost impossible to measure the visual fatigue, thereforeno wonder they approached a way of indirect measurement.Optometric devices, such as autorefractor for measurementof mismatch between accommodation and convergence, havebeen applied in comparisons to determine the amount of changeof accommodation in viewing integral imaging [15], [16]. This

1551-319X/$31.00 © 2012 IEEE

KIM et al.: ACCOMMODATIVE RESPONSE OF INTEGRAL IMAGING IN NEAR DISTANCE 71

approach can objectively analyze the visual fatigue caused bymismatch between accommodation and convergence. However,so far, measurements of an accommodative response of the eyefatigue in integral imaging have never been examined for largesample size of subjects.

In this paper, we present objective evaluation results usingan optometric device to measure accommodative responses inviewing a real object and integral imaging. Measurement datafor the accommodative response of both the real object and thereconstructed image from integral imaging were obtained from71 volunteers (30.9 years of average age, ranging from 23 to 38years, 18% of eyeglasses wearers). The predefined interval ofthe accommodative response was subdivided into 10 cm in neardistance. We performed statistical analysis to compare distinctlyaccommodative response of both cases. By the comparison, wecould verify that more than 73% of participants stared at thereal integrated 3D image from integral imaging instead of thedisplay panel. The data clearly show that accommodation-con-vergence mismatch, which is mentioned as a possible cause forvisual fatigue when watching 3D displays, did not coincide withintegral imaging in near-distance observation.

II. ASSESSMENT OF THE EYE FATIGUE

A. Depth Perception and the Causes Leading to theEye Fatigue

When fixed objects are observed by human eyes, the observeruses both psychological and physiological cues to recognizedepth information of the objects. The psychological cues in-clude a linear perspective, an overlapping, shading, and texturegradient, which could be generated by computer graphics. Thephysiological cues could be classified into an accommodation,a binocular disparity, a convergence, and a motion parallax. Be-cause the two eyes of human are separated horizontally by about6.5 cm and each eye has its own perspective, slightly differentimages are projected to the retina of each eye. The images arecombined by brain, and the depth information is obtained byboth psychological and physiological cues. The difference be-tween a real object and a 3D image based on stereoscopic dis-plays came from artificial depth information. Although a 3Dimage combining the above depth cues is plausible artifacts, itdoes not provide all of depth cues, like real objects. Therefore, itmight give rise to serious visual fatigue such as eyestrain, feelingof pressure in the eyes, an eye ache, a difficulty in focusing, andheadaches [17].

The visual fatigue could be defined as decline of a visionsystem. The causes of visual fatigue are diverse and still highlycontroversial issues for debate. In this paper, we shall focus onnot whole visual fatigue system but the visual fatigue relatedto the 3D display. Visual fatigue when 3D images are observedcould be categorized into many types, such as dry eyes, a diffi-culty in focusing objects, headaches, shoulder pain, nausea, andso on. However, to date, no definitive answer has been given tothe causes leading to these symptoms. Many investigations andresearches concerning causes of the symptoms refer to the mis-match between accommodation and convergence, an excessivebinocular parallax, and a geometrical distortion came from the

Fig. 1. Mismatch between accommodation and convergence in viewing 3D dis-play based on autostereoscopic display.

left and right images. The causes have been elusive; this is at-tributed to intricate complex of visual nerve system and it doesnot appear as a form of resultant visual symptom.

Literature mentions that the mismatch between accommoda-tion and convergence was known as the possible factor for theeye fatigue [18]–[20], but it needs a great deal of debate andresearch to investigate the eye fatigue mechanism. In viewingreal objects, both accommodation and convergence response arecoupled reflexively. It is referred to as a Donder’s line, whichshows the range that is effective in terms of coupled mechanismbetween two stimuli in viewing real scene [21]. On the otherhand, the accommodative stimuli are said to be fixed on stereo-scopic display, while the convergence stimuli vary dependingon the screen disparity as shown in Fig. 1. The mismatch caninduce visual fatigue; in particular, the situation becomes se-rious in near-range observation of 3D images. Depth perceptionin near distance could be affected by binocular disparity, accom-modation, and convergence. The observation of 3D images canbe mostly done within 2 3 m. Therefore, the depth percep-tion in the stereoscopic displays could be strongly affected bythe mismatch inducing the visual fatigue.

Despite relatively vigorous research about visual fatigue inthe field of stereoscopic displays, few have attempted to addressa visual fatigue mechanism in integral imaging. It attracted onlylimited attention; however, it has come to the forefront again inrecent years because of recent progress of the high-resolutioncamera and spatial light modulator (SLM). Integral imaging isa potential candidate to realize 3D displays; however, it is notfree from the visual fatigue as well. Therefore, it is important toinvestigate causes and effects of the visual fatigue in viewing in-tegral imaging. Effective ways to assess the visual fatigue havebeen widely examined in the stereoscopic displays; it could alsobe applied in integral imaging. The following two sections dis-cuss both subjective and objective methods to measure the vi-sual fatigue.

B. Subjective Methods to Assess Eye Fatigue

A classical subjective assessment for visual fatigue is to usequestionnaires as an evaluation indicator. Questionnaires havebeen widely applied to determine the degree of visual fatigue. Asimulator sickness questionnaire (SSQ), which was devised byKennedy et al., is a well-established tool for the investigation of

72 JOURNAL OF DISPLAY TECHNOLOGY, VOL. 8, NO. 2, FEBRUARY 2012

TABLE IQUESTIONNAIRES IN ACCORDANCE WITH EXPERIMENTAL CONDITIONS AND

EVALUATION ITEMS

the visual fatigue [22]. It has been used extensively in studiesof simulator sickness and has 26 possible symptoms and pro-vides scores in four categories: nausea, oculomotor discomfort,disorientation, and general factors. Since then, many modifiedversions of the SSQ have been provided by researchers. How-ever, these questionnaires covered a wide scope because theyincorporated the fields of medical, mental, physiological, andsocial sciences. None of them ever came to standard protocols.Therefore, they are usually modified in accordance with exper-imental conditions and evaluation items. Because the visual fa-tigue is based on subjective experience, these methods can beused in combination with objective assessment. Table I showssubjective methods using questionnaire to assess the visual fa-tigue [9], [23]–[27].

C. Objective Methods to Assess Eye Fatigue

The most effective way to assess the visual fatigue is to findout biological signals indicating the visual fatigue. Because thevisual fatigue is influenced by subjective experiences and it ishard to assess the degree of the visual fatigue, implementing op-tometric devices or medical instruments is one of the potentialcandidates for assessing visual fatigue objectively. These de-vices are applied in measuring characteristic decline of the eyeof physical changes caused by visual fatigue. They could also

TABLE IICLASSIFICATION OF OBJECTIVE METHODS TO ASSESS VISUAL FATIGUE

be used in ‘before and after viewing stereoscopic displays’ testto determine the degree of visual fatigue.

The optometric devices could be classified into two groupsdepending on analytical methods; one is to estimate the visualfatigue using analysis of comprehensive biological reactions.For example, there are twelve cranial nerves, which connect thebrain with the eyes, ears, nose, tongue, etc. When the stimulithat can cause the visual fatigue are perceived, brain activitieshave different perspectives than usual. However, these biolog-ical signals do not indicate visual fatigue directly and have com-plex patterns, therefore it is critical for better understanding vi-sual fatigue to analyze the captured biological reactions. Thesebiological signals can be caught by electrode labeling devices,such as electroencephalography (EEG), magneto-encephalog-raphy (MEG), electromyography (EMG), and electrocardiog-raphy (ECG) [28]–[31]. Numerous researches have been donein a brain activity research field about depth perception basedon stereoscopic displays, and ongoing research has been per-formed. However, it was not exactly revealed that biological sig-nals are associated with causes of the visual fatigue. Another ex-ample to assess the visual fatigue is monitoring certain proteinsin body as a marker for the visual fatigue. Certain proteins, suchas Chromogranin A (CgA) in the saliva, have been reported toincrease with mental or physical stress [32].

The other method to figure out objectively the visual fatigueis to assess direct indicators which appeared in fatigued eye.The optometric devices, such as refractometer and eye tracker,are typical examples of measurement tool. Those deviceswould be categorized as indirect instruments because they donot measure the visual fatigue directly; however, they enableus to assess directly accommodation and eye movements.Hence, they could be classified as a direct assessment method.For pupillary activities, the fatigue or tiredness after viewingstereoscopic images can affect autonomic nervous systems.The assessment of the degree of the visual fatigue has beenconducted on measuring pupil size which is affected by mentalconditions. A refractometer is the most classical method tomeasure accommodative responses of the eye. The device hasbeen used for verifying a hypothesis that predicts differencebetween the before- and after-state in viewing stereoscopicdisplays. Several studies have reported accommodative re-sponses of the eye in viewing stereoscopic displays; howeverthe experiments using optometric devices are time consumingand performed with small number of subjects as shown inTable II [8], [28]–[35].

Recently, the optometric device using open view windowmethod, which adopted the half mirror, made it possible to

KIM et al.: ACCOMMODATIVE RESPONSE OF INTEGRAL IMAGING IN NEAR DISTANCE 73

take a measurement without the subject being aware as thereis no distraction of stereoscopic images. In this paper, usingthe optometric device, we conducted measurement of accom-modative response while each participant is observing integralimaging. The number of participants in the experiment wasmore than 70. Because it never has been clearly proven thatthe accommodative responses have a causal relationship withvisual fatigue, here we assumed that the visual fatigue will berelieved by reducing the mismatch between accommodationand convergence. Although integral imaging can display 3Dimages with full parallax, resulting from a 2D lens array, it hasnever been clarified whether the viewer looks at 3D images or aflat panel display in integral imaging. The purpose of this paperis to verify that the observers’ eyes see the 3D images at desiredpositions as if they saw the real objects in near distance.

III. EXPERIMENTS

Using a modified optometric device (WAM-5500, GrandSeiko Company), we measured the accommodative responseof the eye in the case of both a real object and a reconstructedimage from integral imaging. The validity and repeatabilityof the optometric device was confirmed by several researches[36]. The modified optometric device can make us select anytarget, so it can be a real object or a 3D image based on integralimaging. Additionally, since we can put a target at any distance,we can measure near and far targets as well. A patch was wornover the left or right eye during the experiment. A measurementof accommodative responses of the eye was performed on 71staff and students of Seoul National University Hospital (33male and 38 female), with a mean age of 30.9 years (rangefrom 23 to 38 years, 18% of eyeglasses wearers). The eyesof subjective (left or right) are randomly selected to measurerefractive power of the eye. The collected participants had littleexperience of the 3D display observation. Exclusion criteriaincluded: 1) best-corrected visual acuity of 20/40 or worse ineither eye; 2) myopia or hyperopia of more than 6.00 dioptersin either eye; 3) anisometropia of more than 2.00 diopters; 4)strabismus or nystagmus; and 5) history or presence of otherocular diseases. The experiment followed the tenets of theDeclaration of Helsinki and was approved by the InstitutionalReview Board of Seoul National University Hospital. Informedconsent was obtained from all participants after the details ofthe study were explained. Each subject who participated in theexperiment received comprehensive optometric evaluations ofthe accommodative response. All subjects rested in a comfort-able condition for 3 min before and between experiments.

Before measuring accommodative response in viewing inte-gral imaging, calibrations are necessary to get accurate data.Generally, the diopter is equal to the reciprocal of the focallength measured in meters. However, refractive power of theeye with emmetropia does not show 1.00 diopter when they seeany real target at 1 m because refractive power of the eye isnot the same for everybody. To normalize the diopter unit, weadopted following empirical experiment. The subjects were re-quired to focus on real objects that were placed at intervals of10 cm from 50 cm to 110 cm, and the accommodative responseof the eye was assessed according to the step-wise movement of

Fig. 2. Experimental setup.

a real object. The experiment was repeated five times, attainingthe mean difference, standard deviation, and 95% confidenceintervals. Due to the inherent problems of analyzing cylindricalcomponents in the eyeball’s conventional form, sphere was con-verted into a vector representation: a spherical lens of powermeans spherical equivalence (equal to sphere [cylinder/2]).Total time to perform the measurement was not carried over5 min to avoid distraction of the observers. The measurementwas taken under an ambient illumination of 250 lux for focusingthem straight, which ensured that no readings were rejected be-cause the pupil size was too small to measure the accommoda-tive response [37]. The ophthalmologist reminded the partici-pants explaining what they have to focus on in every step of theexperiment. We used a lozenge pattern that is made of electro-luminescent film (EL film) as shown in Fig. 2. We could assurethe participants saw the target without any hindrance becauseEL film is a self-light-emitting device and the measurement wassignificantly affected when the eccentricity of gaze was over10 from the visual axis. Therefore, an examined eye, modifiedoptometric device, and a real target object were aligned on astraight line to ensure that convergence of the eye was not con-tributing to the measurement. The data were arranged to dif-ferent individuals for comparison with reconstructed 3D imagecase.

The reconstructed image from integral imaging was producedwith following procedures [15], [16]. For simplicity of the ex-periment, we used an overhead projector (OHP) film and printedelemental images on it using a laser printer for elemental im-ages. The printed elemental image resolution is 1200 dpi. TheEL film is behind of the elemental image on OHP film, andlens array is in front of the OHP film. Although the OHP film

74 JOURNAL OF DISPLAY TECHNOLOGY, VOL. 8, NO. 2, FEBRUARY 2012

Fig. 3. Experimental schematic.

TABLE IIIEXPERIMENTAL CONDITIONS

cannot display a moving image or we cannot change the ele-mental image, we can control brightness by applying variablevoltage supply to remove the luminance difference between areal object and a reconstructed 3D image from integral imaging.We called this a display unit. As shown in Fig. 3, the display unitwas fixed 80 cm away from the observer, while altering the el-emental image in order to vary the depth of the reconstructed3D image. The elemental image was displayed so that the re-constructed 3D image is integrated as a real integrated imagelocated 60 and 70 cm from observer. In the case of a virtual in-tegrated image located at 90 and 100 cm, the elemental imagewas adjusted to display the virtual integrated image behind thedisplay unit. Therefore, a real integrated image and a virtual in-tegrated image represent reconstructed 3D images in front andbehind of the display unit by using integral imaging method. Ac-commodative response of the eye was measured five times also,attaining the mean difference, standard deviation, and 95% con-fidence intervals. In the 3D image case, the measurement wasnot carried over 5 min; the experiment was performed on equalterms as the real object case. A detailed description of the ex-perimental setup is shown in Table III.

IV. EXPERIMENTAL RESULTS AND DISCUSSION

Complete recordings for all participants were obtained foreach accommodation stimulus, and accommodative responseswere averaged into five measurements for each accommodationstimulus. Fig. 4 shows sample results of the accommodative re-sponses for real objects. The real target was located at 50, 60,70, 80, 90, 100, and 110 cm, moved with a step response. As

shown in Fig. 4 and Table IV, each subject had different ac-commodative responses when the observers saw a real objectbecause each subject had different refractive power of the eyedespite of strict experimental conditions. For example, whenthe participants were asked to see a real object at a distance of0.5 m (equal to 2.00 diopter), the accommodative responsevaried from 1.80 diopter to 0.60 diopter. This variation re-sulted from different refractive power of the eye of each par-ticipant. Note that it had a tendency of decreasing the absolutevalue of accommodative response as accommodation stimulusincreases. However, absolute value of accommodative responsedid not show exactly the reciprocal of located distance of thereal object. In this case, we cannot figure out where the partici-pants see when they were asked to see the 3D image.

Therefore, we need to obtain the raw data of accommodativeresponse when each subject sees a real object. At first, accom-modative responses of the real object were measured from 50to 110 cm at intervals of 10 cm. After five times repetition ofeach subject, we can obtain accommodative response raw dataof 71 participants when they see the real object. This makes usestimate consecutive accommodative responses of the real ob-ject from 50 to 110 cm per participant, as shown in Fig. 4. Then,we also measured accommodative response of every participantin case of integrated 3D image. The display device was fixed at80 cm from the participant, while the reconstructed 3D imageby integral imaging was formed at 60, 70, 90, and 100 cm.

To probe on the question whether the participants look atthe display panel or not when they were forced to see integralimaging, accommodative responses of the real object and thereconstructed 3D image were compared. Because the displaypanel was fixed at 80 cm when integral imaging was displayed,we compared the accommodative response between 80 cm of areal object case and 60, 70, 90, and 100 cm of the 3D image case.The accommodative response was compared by T-test [38]. TheT-test was executed independently four times per each partici-pant ([60 cm 3D image, 80 cm real object], [70 cm 3D image, 80cm real object], [90 cm 3D image, 80 cm real object], and [100cm 3D image, 80 cm real object]), and we performed a T-testper participant independently (Results were interpreted sepa-rately as statistically significant when values were 0.05).However, because it was not confirmed whether measurementsamples have same standard deviation or not, we carried outF-test before independent T-test and Welch’s corrected T-testwere conducted [39]. When two groups have the same standarddeviation distribution, independent T-test was carried out afterthe F-test was conducted to each sample. The Welch’s correctedT-test was conducted when they have a different standard devi-ation distribution.

Fig. 5 shows the comparison of accommodative responseswhen they saw a real object at the position of 80 cm and a 3Dimage of each case. If the accommodative response of the partic-ipant in viewing a real object was different from that in viewingan integrated 3D image, we marked it “different.” It representedthey did not see the display unit. As shown in Fig. 5, more than62% of the participants (44 participants) did not focus on thedisplay unit when they saw the integrated 3D image (both areal integrated 3D image and a virtual integrated 3D image). Inother words, the participants did not keep eyes on display unit

KIM et al.: ACCOMMODATIVE RESPONSE OF INTEGRAL IMAGING IN NEAR DISTANCE 75

Fig. 4. Accommodative response measurements in viewing a real object (sample of six participants).

TABLE IVACCOMMODATIVE RESPONSE MEASUREMENTS IN VIEWING A REAL OBJECT

FOR SIX SUBJECTS IN OPTOMETRIC EVALUATION TESTS (STD: STANDARD

DEVIATION, CL: CONTROL LIMIT)

in near-distance (within 1 m) when they saw an integrated 3Dimage. Moreover, the result indicates that participants had moredistinctive accommodative response (over 73%, 52 participants)in viewing real integrated 3D images (integrated at the positionof 60 and 70 cm) than in viewing virtual integrated 3D images(integrated at the position of 90 and 100 cm).

However, in the above analysis, it was not clarified where theparticipants exactly see in viewing integral imaging. Because wemeasured accommodative responses of each participant from 50to 110 cm of a real object at intervals of 10 cm, the accommoda-tive response curve in near distance could be acquired when they

Fig. 5. Participant numbers that have different accommodative response froma real object at the position of 80 cm when they saw a 3D image (integrated 3Dimages at the position of 60, 70, 90, and 100 cm).

saw the real object. In order to estimate the concordance of thereal object and the integrated 3D image, we compared accom-modative responses at each step in the integrated 3D image andaccommodation curve of the real object. Fig. 6 shows the re-sults. When the participants saw the real integrated 3D image at60 cm, more than 65% of participants recognized it as seeinga real object at a range of 50 70 cm (50 60 cm: about30%, 60 70 cm: about 35% of concordance). Also, the re-sults show a similar tendency at the 70 cm of a real integratedimage (50 60 cm: about 25%, 60 70 cm: about 35% of con-cordance). On the contrary, when the virtual integrated imagewas displayed behind of the display unit at 90 and 100 cm, theyrecognized it as if it were located at the whole range of exper-iment although more than 62% of participants did not see thedisplay panel itself as shown in Fig. 5.

Given these statistical results we could take into account twoimportant things. The result reflects that the participants kepttheir eyes on the real integrated 3D image when the recon-structed 3D image was displayed in front of the display unit.Secondly, when the reconstructed virtual 3D image was dis-played behind of the display unit, the proportion distributionof accommodative response of participants was concentrated

76 JOURNAL OF DISPLAY TECHNOLOGY, VOL. 8, NO. 2, FEBRUARY 2012

Fig. 6. Accommodative response distribution of participants when they see theintegrated 3D image (real and virtual case).

on the central range (60 90 cm). This result shows a dif-ferent tendency from the case of a real integrated 3D image.From the accommodative response measurements of real andvirtual integrated images, we can deduce that the participantsfocused their eyes on the real integrated 3D image than vir-tual integrated 3D image. In the beginning of the experiment,we assumed that visual fatigue can be relieved when the par-ticipants focus on the integrated 3D image. Based on this as-sumption, we can deduce that the participants were less sensitiveon visual fatigue caused by mismatch between accommodationand convergence when they saw the real integrated 3D image innear-distance. However, in the virtual integrated 3D image case,the accommodation of people is widely distributed between 70cm and 90 cm. Here we need to point out that we are usingcentimeter unit. The use of centimeter unit instead of diopterunit makes accommodative response less sensitive, especiallywhen the integrated 3D image was formed behind of displayunit. For instance, 50 60 cm could be transformed to dioptervalue of 1.67 2.00, while 100 110 cm could be transformedto 0.91 1.00 diopter. Given the accommodative response of thehuman eye reacts linearly to the diopter unit, the interval of10 cm was not a distinct difference as the accommodative re-sponses were measured far from the participants [40]. It couldbe seen as the reason of the peculiar distribution at the range of100 110 cm. Another reason that the accommodative responseseems to be widely distributed in the case of virtual 3D imagecould be the contribution of monocular eye according to the dis-tance. Although integral imaging used a lens array method likea lenticular lens method, it is still debatable whether monoc-ular eye or binocular eyes are effective when integral imagingis displayed. In the experiments, the accommodative responsewas measured by monocular eye which is an effective depthcue within 2 m [7]. Therefore, additional experiments will berequired to conclude that participants accommodate at the po-sition where the virtual 3D image is displayed. The range ofrendered image will be suitable in 1 2 m.

V. CONCLUSION

Accommodative response of integral imaging in near distancefor adult subjects was obtained. Despite difficulties such as ex-panding large sample size of participants, subdivision of the3D image interval and time consuming experiment, seventy-one

normal adult subjects participated in the experiments. The ex-periment was performed within 1 m, and the interval was subdi-vided into 10 cm. The accommodative response was measuredin both real and virtual integrated images. The presented resultsshowed that over 73% of participants kept eyes on the real in-tegrated 3D image instead of display panel, while the propor-tion distribution of accommodative response of participants waswidely distributed on the central range in the case of virtual in-tegrated 3D image. Based on the assumption that participantssuffered less visual fatigue by reducing the mismatch betweenaccommodation and convergence, we concluded that visual fa-tigue will be relieved when they accommodated at the positionwhere the 3D integrated image was displayed in front of the dis-play panel. In the virtual integrated 3D image case, additionalexperiments within monocular perception limit will be required.

REFERENCES

[1] W. Matusik and H. Pfister, “3D TV: A scalable system for real-timeacquisition, transmission, and autostereoscopic display of dynamicscenes,” ACM Trans. Graphics, vol. 23, pp. 814–824, 2004.

[2] B. Lee, J.-H. Park, and S.-W. Min, “Three-dimensional display andinformation processing based on integral imaging,” in Digital Holog-raphy and Three-Dimensional Display, T.-C. Poon, Ed. New York:Springer, 2006, ch. 12.

[3] J.-H. Park, K. Hong, and B. Lee, “Recent progress in three-dimensionalinformation processing based on integral imaging,” Appl. Opt., vol. 48,pp. H77–H94, 2009.

[4] B. Javidi and F. Okano, Three Dimensional Television, Video, and Dis-play Technology. Berlin, Germany: Springer, 2002.

[5] T.-C. Poon, “Optical scanning holography—A review of recentprogress,” J. Opt. Soc. Korea, vol. 13, pp. 406–415, 2009.

[6] A. Stern and B. Javidi, “3-D computational synthetic aperture integralimaging (COMPSAII),” Opt. Express, vol. 11, pp. 2446–2451, 2003.

[7] M. Lambooij, W. IJsselsteijn, M. Fortuin, and I. Heynderickx, “Visualdiscomfort and visual fatigue of stereoscopic displays: A review,” J.Imaging Sci. Technol., vol. 53, pp. 030201–1–030201-14, 2009.

[8] T. Inoue and H. Ohzu, “Accommodation responses to stereoscopicthree-dimensional display,” Appl. Opt., vol. 36, pp. 4509–4515, 1997.

[9] S. Yano, S. Ide, T. Mitsuhashi, and H. Thwaites, “A study of visualfatigue and visual comfort for 3D HDTV/HDTV images,” Displays,vol. 23, pp. 191–201, 2002.

[10] Y. Okada, K. Ukai, J. S. Wolffsohn, B. Gilbartin, A. Iijima, and T.Bando, “Target spatial frequency determines the response to conflictingdefocus- and convergence-driven accommodative stimuli,” Vision Res.,vol. 46, pp. 475–484, 2006.

[11] S.-K. Kim, D.-W. Kim, Y. M. Kwon, and J. Y. Son, “Evaluation ofthe monocular depth cue in 3D displays,” Opt. Express, vol. 16, pp.21415–21422, 2008.

[12] J. Kim, H. K. Yang, Y. Kim, B. Lee, and J.-M. Hwang, “Distancestereotest using a 3D monitor for adult subjects,” Amer. J. Ophthalmol.,vol. 151, pp. 1081–1086, 2011.

[13] G. D. Love, D. M. Hoffman, P. J. W. Hands, J. Gao, A. K. Kirby, andM. S. Banks, “High-speed switchable lens enables the development of avolumetric stereoscopic display,” Opt. Exp., vol. 17, pp. 15716–15725,2009.

[14] H. Nagatani and Y. Hirayama, “Evaluation of the influence on thehuman body of the autostereoscopic display based on integral imagingmethod,” in Proc. SPIE, 2008, vol. 6803, pp. 68030E-1–68030E-8.

[15] Y. Kim, J.-H. Jung, K. Hong, G. Park, and B. Lee, “Accommodationresponse in viewing integral imaging,” in SID Int. Symp. Digest Tech.Papers, 2010, pp. 530–532.

[16] Y. Kim, K. Hong, J. Kim, H. K. Yang, J.-M. Hwang, and B. Lee, “Ac-commodation response measurement according to angular resolutiondensity in three-dimensional display,” in Proc. SPIE, 2011, vol. 7956,pp. 79560Q–79560Q-7.

[17] I. P. Howard and B. J. Rogers, Binocular Vision and Stereopsis. Ox-ford, U.K.: Oxford Univ., 1995.

[18] A. Wilkins, Visual Stress. Oxford, U.K.: Oxford Univ., 1995.[19] D. M. Hoffman, A. R. Girshick, K. Akeley, and M. S. Banks, “Ver-

gence-accommodation conflicts hinder visual performance and causevisual fatigue,” J. Vision, vol. 8, pp. 1–30, 2008.

KIM et al.: ACCOMMODATIVE RESPONSE OF INTEGRAL IMAGING IN NEAR DISTANCE 77

[20] M. Emoto, T. Niida, and F. Okano, “Repeated vergence adaptationcauses the decline of visual functions in watching stereoscopic tele-vision,” J. Display Technol., vol. 1, pp. 328–340, 2005.

[21] I. M. Irvin, Clinical Refraction, 3rd. ed. New York: ProfessionalPress, 1970.

[22] R. S. Kennedy, N. E. Lane, K. S. Berbaum, and M. G. Lilienthal, “Sim-ulator sickness questionnaire: An enhanced method for quantifyingsimulator sickness,” Int. J. Aviat. Psychol., vol. 3, pp. 203–220, 1993.

[23] J. Kuze and K. Ukai, “Subjective evaluation of visual fatigue causedby motion images,” Displays, vol. 29, pp. 159–166, 2008.

[24] S. L. Ames, J. S. Wolffsohn, and N. A. Mcbrien, “The developmentof a symptom questionnaire for assessing virtual reality viewing usinga head-mounted display,” Optometry Vis. Sci., vol. 82, pp. 168–176,2005.

[25] M. Ogata, K. Ukai, and T. Kawai, “Visual fatigue in congenital nys-tagmus caused by viewing images of color sequential projectors,” J.Display Technol., vol. 1, pp. 314–320, 2005.

[26] K. Murata, S. Araki, N. Kawakami, Y. Saito, and E. Hino, “Centralnervous system effects and visual fatigue in VDT workers,” Int. Arch.Occup. Environ. Health, vol. 63, pp. 109–113, 1991.

[27] J. Häkkinen, M. Pölönen, J. Takatalo, and G. Nyman, “Simulator sick-ness in virtual display gaming: A comparison of stereoscopic and non-stereoscopic situations,” in Proc. 8th Conf. Human-Comput. Interac-tion With Mobile Devices and Services, New York, 2006, pp. 227–230.

[28] K. Murata, S. Araki, K. Yokoyama, K. Yamashita, T. Okumatsu, and S.Sakou, “Accumulation of VDT work-related visual fatigue assessed byvisual evoked potential, near point distance and critical flicker fusion,”Industrial Health, vol. 34, pp. 61–69, 1996.

[29] S. Yamamoto and S. Matsuoka, “Topographic EEG study of visual dis-play terminal (VDT) performance with special reference to frontal mid-line theta waves,” Brain Topogr., vol. 2, pp. 257–267, 1990.

[30] B. T. Jap, S. Lal, P. Fischer, and E. Bekiaris, “Using EEG spectralcomponents to assess algorithms for detecting fatigue,” Expert Syst.Applic., vol. 36, pp. 2352–2359, 2009.

[31] A. Zhang, Z. Zhao, and Y. Wang, “Estimating VDT visual fatiguebased on the features of ECG waveform,” in Proc. Int. Workshop Inf.Security Applic., F. Gao and X. Zhu, Eds., Oulu, 2009, pp. 446–449.

[32] S. Fujimoto, M. Nomura, M. Niki, H. Motoba, K. Ieishi, T. Mori, H.Ikefuji, and S. Ito, “Evaluation of stress reactions during upper gas-trointestinal endoscopy in elderly patients: Assessment of mental stressusing chromogranin A,” J. Med. Invest, vol. 54, pp. 140–145, 2007.

[33] T. Shibata, T. Kawai, K. Ohta, M. Otsuki, N. Miyake, Y. Yoshihara,and T. Iwasaki, “Stereoscopic 3-D display with optical correction forthe reduction of the discrepancy between accommodation and conver-gence,” J. Soc. Inf. Disp., vol. 13, pp. 665–671, 2005.

[34] M. Miyao, S. S. Hacisalihzade, J. S. Allen, and L. W. Stark, “Effectsof VDT resolution on visual fatigue and readability: An eye movementapproach,” Ergonomics, vol. 32, pp. 603–614, 1989.

[35] H. T. Nguyen, D. M. Isaacowitz, and P. A. D. Rubin, “Age- and fa-tigue-related markers of human faces: An eye-tracking study,” Oph-thalmology, vol. 116, pp. 355–360, 2009.

[36] A. L. Sheppard and L. N. Davies, “Clinical evaluation of the GrandSeiko auto ref/keratometer WAM-5500,” Ophthalmic Physiol. Opt.,vol. 30, pp. 143–151, 2010.

[37] M. S. Livingstone and D. H. Hubel, “Stereopsis and positional acuityunder dark adaptation,” Vision Res., vol. 34, pp. 799–802, 1994.

[38] B. L. welch, “The significance of the difference between two meanswhen the population variances are unequal,” Biometrika, vol. 29, pp.350–362, 1937.

[39] D. J. Best and J. C. W. Raynor, “Welch’s approximate solution for theBehrens-Fisher problem,” Technometrics, vol. 29, pp. 205–210, 1987.

[40] O. Franzén, H. Richter, and L. Stark, Accommodation and VergenceMechanisms in the Visual System. New York: Birkhäuser Verlag,2000.

Youngmin Kim received the B.S. degree and Ph.D.degree in electrical engineering from Seoul NationalUniversity, Seoul, Korea, in 2005 and 2011, respec-tively.

He is currently with Seoul National University asa Postdoctoral Researcher. His current research inter-ests focus on the 3-D display, image processing in-cluding multiview 3D display, and visual fatigue ofhuman vision related on 3D display.

Jongshin Kim received the M.D. degree from theCollege of Medicine, Seoul National University,Seoul, Korea, in 2006. He is currently workingtoward the Ph.D. degree in mechanical engineeringfrom KAIST, Seoul, Korea.

After a general internship, he completed his resi-dency in Ophthalmology at Seoul National Univer-sity Hospital in February 2011. His research interestsfocus on development of new biomedical optical in-strumentation for high-resolution in vivo retina andchoroid imaging (allowing visualization of individual

cellular structures).

Keehoon Hong received the B.S. degree in electricalengineering from Yonsei University, Seoul, Korea, in2008. He is currently working toward the Ph.D. de-gree at the School of Electrical Engineering, SeoulNational University, Seoul, Korea.

His primary research interest is in the areas of 3Ddisplay and 3D image processing.

Hee Kyung Yang received the M.D. degree fromthe College of Medicine, Seoul National University,Seoul, Korea, in 2004.

She is a Clinical and Research Fellow with the De-partment of Ophthalmology, Seoul National Univer-sity Bundang Hospital, Seoul, Korea. She completedan internship and residency in Seoul National Univer-sity Hospital in 2009. Her current research of inter-ests includes fields of pediatric ophthalmology, stra-bismus and neuro-ophthalmology.

Jae-Hyun Jung received the B.S. degree in elec-tronics engineering from Pusan National University,Pusan, Korea, in 2007. He is currently workingtoward the Ph.D. degree in electrical engineering atSeoul National University, Seoul, Korea.

His primary research interests are in the areas of3D display, optical information processing, and depthperception in human visual system.

Heejin Choi received the Ph.D. degree in electricalengineering from Seoul National University, Seoul,Korea, in 2008.

After working for Samsung Electronics, in 2010,he joined the Department of Physics, Sejong Univer-sity, Seoul, Korea. His primary research interest is inthe areas of 3D displays, 3D/2D conversion, and 3Dhuman factor.

78 JOURNAL OF DISPLAY TECHNOLOGY, VOL. 8, NO. 2, FEBRUARY 2012

Sung-Wook Min received the B.S. and M.S. degreesin electrical engineering and the Ph.D. degree fromSeoul National University, Seoul, Korea, in 1995,1997, and 2004, respectively.

Presently, he is a faculty member with the Depart-ment of Information Display, Kyung Hee University,Seoul, Korea, which he joined in 2007. He is in-terested in 3D imaging and the advanced displaysystem, especially based on the integral imagingtechnique.

Jong-Mo Seo received the M.D. degree and M.S. andPh.D. degrees in biomedical engineering from SeoulNational University (SNU) School of Medicine,Seoul, Korea, in 1996, 2002, and 2005 respectively.

He completed the ophthalmologic residency in2002 and the retina fellowship in 2005 at SNUHospital (SNUH). He served as a Clinical Instructorwith the Department of Ophthalmology, SNUH, and,since 2008, he has been an Assistant Professor withthe Department of Electrical Engineering at SNUSchool of Engineering. His research interest is the

artificial retina for the blind, neural prosthesis, visual perception mechanism,and medical information technologies.

Jeoung-Min Hwang received the M.D. degree fromSeoul National University College of Medicine,Seoul National University, Seoul, Korea, in 1985.

She is a Professor with the Department of Oph-thalmology, Seoul National University BundangHospital, Seoul, Korea. Her current research ofinterests includes fields of pediatric ophthalmology,strabismus, and neuro-ophthalmology.

Byoungho Lee received the Ph.D. degree in electricalengineering and computer science from the Univer-sity of California, Berkeley, in 1993.

Since 1994, he has been with the School of Elec-trical Engineering, Seoul National University, Seoul,Korea, as a faculty member, where he is currentlya Full Professor. His research group has publishedmore than 280 international journal papers and pre-sented more than 490 international conference papersincluding more than 90 invited papers. His researchfields are 3D display, diffractive optics, fiber devices,

and surface plasmon–polariton applications.Dr. Lee is a Fellow of the Optical Society of America and the International

Society for Optical Engineers. He received the 5th Young Presidential Scien-tist Award of Korea in 2002 and the Scientist of the Month Award of Korea inSeptember 2009.