鈴木久貴 1)，鈴木百合彩 2)，津田良太郎 2)，Vimuktalo

鈴木久貴 1)，鈴木百合彩 2)，津田良太郎 2)，Vimuktalop Pinyapach3)，Jaruthien Nine3)，白井暁彦 4)
1) 神奈川工科大学 情報工学専攻（〒 243-0292 神奈川県厚木市下荻野 1030, [e-mail]） 2) 神奈川工科大学 情報学部 情報メディア学科 3) Chulalongkorn University, Department of Imaging and Printing Technology（254 Phyathai Road, Patumwan, Bangkok Thailand. 10330） 4) 神奈川工科大学 情報学部 情報メディア学科 准教授




概要: 家庭用ゲーム機など，さまざまな HDMI 入力を多重化できる FPGA ハードウェア化により，よ り幅広い応用の可能性が広がった多重化不可視映像技術「ExPixel」の一般化に向け，普及型表示デバ イスにおける視聴特性とその評価を行った．視聴者のディスプレイとの距離および，ディスプレイに対 する角度によるディスプレイ輝度の変化について，円偏光フィルタを用いた場合で実験を行った．また 各ディスプレイメーカーから発売されている，既存のパッシブ式 3D ディスプレイの映像設定について， デフォルトの設定と γ = 2.2 にキャリブレーションされたときのディスプレイ輝度の変化について実験 を行った結果を報告する． キーワード： 多重化，偏光，3D ディスプレイ




1. はじめに 家庭用の 3D ディスプレイが一般化してきたが，立体感 を上手く表現できているコンテンツの供給はハードウェア の供給に追いついていないのが現状である．歴史を振り返 ると，3D 立体映像は一過性のブームを呼ぶ傾向があり 2～ 4 年で鎮静化する傾向にある [1]．テレビ市場では 3D ディ スプレイに代わって，4K 解像度を売りにする製品が増えた ことを考えると，今回のブームも鎮静化に傾いていると認 めざるを得ない．このような背景から，我々は民生品の 3D ディスプレイに新たな付加価値を与える技術として，多重 化不可視映像技術「ExPixel」を提案している [2]．本技術 は民生品パッシブ式 3D フラットパネルにおいての円偏光と 偶数奇数ピクセル列を利用して，円偏光フィルタ装着者と 裸眼視聴者に対して全く別の映像を提示できる技術である． 研究当初は GPU ベースで動作するシェーダー技術として 多重化不可視映像を実現していたが，近年では FPGA ハー ドウェア化にも成功したことにより [3]，既存のコンシュー マーコンテンツの多重化が可能となった．本報告では民生 品 3D ディスプレイにおいて多重化不可視映像を視聴する 上で適切な視聴領域の調査，多様な 3D ディスプレイにおけ る多重化不可視映像の良好な視聴のために，普及型パッシ ブ式 3D ディスプレイの視聴特性について実験した結果を 報告するものである．
Hisataka SUZUKI, Yuria SUZUKI, Ryotaro TSUDA, Vimuktalop Pinyapach, Jaruthien Nine, and Akihiko SHIRAI




2. 関連研究 多重化映像技術は多くの研究者によって試みられている． Mistry による「ThirdEye」[4] は CRT，LCD プロジェク タといったスクリーンに同期して120Hz以上の速度でシャッ ターの開閉を行うメガネを用いて多重シナリオ上映の提案 を行った．このような高速シャッターを用いた時分割によ る方法は Fakespace Labs 社 McDowall らによる「Snared Illumination」[5]および Multi Use Light Engine (MULE) プロジェクタによって既に提案されている．「ThirdEye」で は裸眼の視聴者に 2 つの映像が重なって見えてしまい隠蔽 が行えていないが，「Snared Illumination」では 2 つの映 像コンテンツの投射回数に十分な差を持たせることで，ア クティブシャッター方式における不可視映像技術を実現し ている．これらと同様のアクティブシャッターを用いたマ ルチプレイヤーゲーム技術が 2012 年米国 Sony Computer Entertainment 社より「SimulView」として特許取得され ているがこちらも両チャネルが多重多重表示される [6]． また，通常の液晶フラットパネルディスプレイの特性調 査は山本らにより 2007 年に報告されている [7]．一方で山 本は薄型ディスプレイの特性・画質評価技術についても報 告している [8]．



3. 実験 まず ExPixel による多重化コンテンツを視聴するための 最適視聴領域の調査，および多様な民生品 3D ディスプレイ において，ExPixel コンテンツの最適な表示を得るために
映像設定評価を行った． 3.1 最適視聴領域の調査 最適視聴領域の調査のために以下にの 3 実験を行った． ExPixel で必要となるパッシブ式 3D ディスプレイデバイス 各種において，フィルタの有無とともに，実験 1) ディスプ レイと観察者の距離 (L)，実験 2) ディスプレイに対する観 察者の水平位置 (角度 θ)，実験 3) ディスプレイに対する観 察者の垂直位置 (角度 θ) の，3 つの実験についてそれぞれ の位置で視聴した際のディスプレイの輝度を測定した．ディ スプレイ上の測定位置は [8] に倣い，画面中央の一定領域 とする．今回の実験ではディスプレイ中央にマスクを施し， 25cm2 内の領域を，デジタル一眼 NikonD750 にて撮影し ディスプレイの輝度の代表値として用いた．位置およびディ スプレイの反射により，自然光による誤差が生じるため実 験環境は暗室とした．ディスプレイに対しカメラを三脚に設 置し，輝度を測りたい各位置にて画像を撮影した．図 1 は 直線距離の測定環境を示しており，ディスプレイ中央の窓に 対して直交するよう 1000mm から 2400mm の間を 200mm 間隔で，全 8 地点で撮影を行った．
図 1: 距離による輝度値の変化の確認
実験 2 について，図 2 は水平方向の視聴位置変化時の輝 度変化を調査したときの測定環境を示している．一般的な ディスプレイの最適視聴距離といわれる 3 ×ディスプレイ の高さ (H) の距離である，1600mm 上で θ = ±60◦ となる ように計 12 点の位置で円偏光フィルタの有無で撮影した．
図 2: 水平角 θ に対する輝度値の変化の実験
垂直方向の調査ではディスプレイを 90◦ 回転させたうえ
で図 2 の方法と同じ位置で θ = ±30◦ がつくように計 6 点 の位置で撮影した．このとき距離および角度を求めるため （株）サムライマーケティング社のレーザー距離計MK-LAK を用いた．角度の算出には本デバイスの機能である，ピタ ゴラスの定理モードを用いた． 輝度値 (luminosity) は，ディスプレイの中央 25cm2 の 領域をカメラで撮影し，その画像の各ピクセルの RGB 各 色の値を加算してそのピクセルの輝度値とした．データに 用いた輝度値は対象の領域全体のピクセル輝度値を加算し て母数で割った平均値を用いた．今回の実験では偏光式 3D ディスプレイである Toshiba Regza42Z8 を用いた． また撮影機材であるカメラは NikonD750 を用いた．予備 実験として，シャッタースピードを｛1/500，1/250，1/125， 1/60，1/30，1/15，1/8，1/4，1/2｝で試し，各シャッター スピード間の差が一定の範囲に収まる 1/30 を基準値とし た．予備実験より，1/250 や 1/8 より外の値では画像が著 しく黒や白になってしまっていた．撮影時の F 値は F/16， ISO は 400 で統一した．カメラとディスプレイの間に円偏 光フィルタを挟まない撮影パターンと，円偏光フィルタを 挟む撮影パターンの 2 パターンを撮影した． 3.2 ディスプレイ特性の評価 現在でもディスプレイ市場では 4K 解像度対応ディスプ レイが普及しており，3D は機能の一部として当たり前のよ うに組み込まれるようになったが，今後もこの勢いは加速 の一途をたどるだろう．本節では現在手に入る民生品の 3D ディスプレイから表 1 に示す機種を使い，明るさおよびコ ントラストの設定を中心として，ExPixel コンテンツの視 聴に適した設定を調査した結果を述べる．
表 1: 対象とした 3D ディスプレイ型名
No. ディスプレイ型名
1 東芝 REGZA42Z8
2 SONY BRAVIA KDL-42W800B
3 LG Smart TV 42LB6700
本調査ではまず表 1 の各製品を工場出荷状態の設定に戻 し，その状態を “キャリブレーション前” 状態とした．まず キャリブレーション前状態の各ディスプレイにおいて “円 偏光フィルタを無使用（フィルタ無し）” の場合，“円偏光 フィルタを使用（フィルタ有り）” の場合の 2 パターンで 調査を行った．その後，各ディスプレイのキャリブレーショ ンを datacolor 社の Spyder4Elite を用いて行い，すべての ディスプレイを “キャリブレーション” 状態とした．同様に 各ディスプレイにおいて “円偏光フィルターを無使用” の’ 場合，“円偏光フィルターを使用” の場合の 2 パターンで調 査を行い，各状態の輝度値を比較した．

4. 結果 4.1 最適視聴領域の特定 4.1.1 距離が輝度値に与える影響の実験 ディスプレイと観察者との間の距離による輝度値の変化 について調査した．通常ディスプレイ輝度は ISO によって 求めるが，本実験においては得られた結果を幅広いディス プレイで利用するため，一眼カメラで撮影した画像から対 象のピクセル値を Luminosity(level) として利用した．図 1 の環境にてディスプレイ視聴者間の距離とディスプレイ輝 度の結果を図 3 に示す．
図 3: 距離-輝度（円偏光フィルタ無し）
図3から，ディスプレイ-カメラ間距離1000mmから2400mm の間において，急激な変化がなくほぼ一直線のグラフになっ ている．これは，通常の輝度と照度の関係では点光源の放 射として考えられるのに対し，普及しているディスプレイ の多くは，バックライトの指向性が制御されており，平行光 源として振る舞う事を示している．(そのためこのような手 法で測定している．) 図 4 は偏光フィルタをカメラのレンズに装着したときの 同条件における輝度と距離の関係を示している．
図 4: 距離-輝度（円偏光フィルタ有り）
フィルタ有りの場合も，全体的な輝度は落ちているが距 離による輝度の極端な低下は見られなかった．全体的に値 が，図 3 の半分に落ちていることから，円偏光フィルタに よって遮断されたピクセル列の光量が落ちたことがわかる． 現象としては輝度は落ちたが，円偏光フィルタに対応する ピクセル列の遮断による撮影画像上の輝度が落ちたため (図 5) であり，多重化不可視映像で広く使われるパッシブ型円
偏光フィルタはディスプレイ強度を減衰させる要因にはなっ ておらず，距離の影響も受けない．
図 5: 円偏光フィルタに対応した偏光のピクセル列が遮断さ れている
4.1.2 水平角度による輝度値の実験 ディスプレイに対する観察者の水平方向の位置の変化に よる輝度の変化を調査した．一般的なディスプレイの最適 視聴距離といわれる 3 ×ディスプレイの高さ (H) の距離で ある，1600mm 上で角度±60◦ がつくように計 12 点の位置 で撮影した (図 2)． 図 6 からディスプレイに対して真正面から，角度がつい ていくことで輝度が下がっていることが言える．
図 6: 水平角 θ-輝度（円偏光フィルタ無し）
図 7 から円偏光フィルタを通したときも距離の実験と同 様，全体の輝度は落ちているが，グラフの傾向としては円 偏光フィルタを通さないときと同じである．
図 7: 水平角 θ-輝度（円偏光フィルタ有り）
4.1.3 垂直角度による輝度値の実験 ディスプレイに対する観察者の垂直角 θ の変化による輝 度の変化を調査した．この調査ではディスプレイを 90◦ 回 転させたうえで図 2 の方法と同じ位置で θ = ±30◦ となる
よう計 6 点の位置で撮影した．垂直方向に角度がついた場 合においても水平方向同様，角度がつくことによる輝度値 の減少が確認できた（図 8，図 9）．
図 8: 垂直角 θ-輝度（円偏光フィルタ無し）
図 9: 垂直角 θ-輝度（円偏光フィルタ有り）
4.2 ディスプレイ毎の輝度値の比較 既存の普及型パッシブ式 3D フラットパネルについて，デ フォルトの映像の輝度値と，ExPixel においてターゲット としている γ = 2.2 にキャリブレーションした設定の輝度 値の変化を比べた．図 10 はキャリブレーションの前後で円 偏光フィルタを通した際の，輝度値の値を表している．こ の値の算出方法は視聴範囲の実験時に用いた方法と同じく， 各ピクセル輝度の値の平均を取った．
図 10: キャリブレーション前後における輝度値の変化
図 10 から各会社のディスプレイはデフォルト状態では， 高い輝度値に設定されてお

ผลลัพธ์ (อังกฤษ) 1:

Suzuki Hisataka 1), Yuri Aya 2 Suzuki), Tsuda, Ryotaro 2), Vimuktalop Pinyapach3), Jaruthien Nine3), Akihiko Shirai 4) 1) Kanagawa Institute of Technology University of Information Engineering (243-0292, Kanagawa, Atsugi-Shi under Ogino 1030, Hisatakaattoshirai.La) 2) Kanagawa Institute of Technology University Faculty of Information Media, Department of 3) Chulalongkorn University, Department of Imaging and Printing Technology (254 Phyathai Road, Patumwan, Bangkok Thailand. 10330) 4) Associate Professor , Kanagawa Institute of Technology University Department of Information Media Overview: Different HDMI inputs, SUCH as Game Consoles, CAN BE multiplexed FPGA hardware Improved, I for generalized spread the Possibility of Wider Application multiplexing Invisible image Technology "ExPixel", spread type device place Audience Characteristics and its Evaluation Conducted. Viewers display Distance and on the display if you Change display brightness angle by USING the circular POLARIZING filter Experiments Were Carried out. Also to report experimental results HAS Been Released from the display, calibrated gamma = 2.2 as the default setting for passive 3D display an existing Visual Settings, When the display brightness changes. Keywords: multiplexing, polarized 3D display1. supply of introduction HAS Been generalized for home 3D display, three-dimensional sense of well Being able to Express content IS current hardware supply does not Keep with. History, Recalls, and 3D stereoscopic Images TEND to FAD to call 2 ~ IS the Tendency to sedation in Four years [1]. Think Increases in the Television Market on behalf of the 3D display, 4 K Resolution to Sell Products, and this Boom Also leans to Quieting down and do not Get floods. From this background, we Give new value-Added Consumer Products on 3D display Technology proposed, Invisible image Technology "ExPixel" [2] ... This Technology IS Present by USING circularly polarized light in Consumer Products passive expression 3D flat panel and even an odd pixel columns, circular Polarizer filter wearers and viewers naked of an Entirely Different Technology. By shader Technology Research INITIALLY Operates in GPU-based multiplexing for Invisible Images WAS That, in recent years successful FPGA hardware for multiplex 3, existing Consumer Mar content Became Possible. In Keep adequate space for Audience Research, various 3D display on paper to Watch multiplex Invisible image in 3D Consumer Products, good viewers Are multiplexed Invisible Images to viewers of mass-Market 30Years drive 3D display full experimental results to report. SUZUKI Hisataka, Yuria SUZUKI, Ryotaro TSUDA , Vimuktalop Pinyapach, Jaruthien Nine, and Akihiko SHIRAI 2. related Research multiplexing of video Technology IS Considered by many Researchers. By Mistry "ThirdEye", [4] Carried out multiple Scenarios Showing proposed glasses sync with CRT, LCD projector, screen, make opening and closing the shutter at A rate of 120 Hz or more. At SUCH high speed shutter by USING time-resolved by the way "Snared Illumination" by Fakespace Labs McDowall et al., [5] and Multi Use Light Engine (MULE) projector already proposed. Delivers Invisible image Technology in A Active shutter system, make enough difference to the Projection of video content of the Two in "Snared Illumination" "ThirdEye" Covered by Footage of Two naked Audience as Unhygienic and hiding but we Have not, but Are you ... Ma lucid player Game Technology USING Active shutter Similar 2012 than United States Sony Computer Entertainment Inc. as A "SimulView" patent and That Multiplex Multiplex Appear both channels but Also [6]. Also, ordinary LCD flat Characteristics survey in 2007 Have Been Reported by Yamamoto et al. [7] ... On the Other hand Mountain's book [8] Have Been Reported on the Characteristics of flat-panel display and video quality Assessment Technologies.3. Experiment First ExPixel investigation of the optimal viewing area for viewing multiplexing content with, and in a variety of consumer products 3D display, in order to obtain an optimal display of the ExPixel content video settings were evaluated... 3.1 were subjected to the following in three experiments to investigate the investigation optimal viewing area of the optimum viewing area. In the passive-type 3D display device various required by ExPixel, along with the presence or absence of a filter, Experiment 1) Disupu Ray and the observer's distance of (L), Experiment 2) horizontal position of the observation person with respect to the display (angle θ), Experiment 3) the vertical position of the observation person for display (the angle θ), to measure the brightness of the display when it is viewed at each position for the three experiments. Measurement position on the di splay copying in [8], and the center of the screen certain areas... In this experiment is masked in the display center, in an area 25 cm 2, which was used as a representative value of the luminance of the captured display by a digital single-lens NikonD750... By the reflection of the position and displays, experimental environment for error due to the natural light is generated was a darkroom... The camera is for Installation on a tripod to display and capture an image at each position wants to measure the brightness... Figure 1 shows a measurement environment of the linear distance, in 200 mm 2400 mm intervals between from 1000 mm to ortcogonally against the center of the display window, and were taken in all eight locations...Figure 1: Distance confirm the change in the luminance value due to the experiment 2, Fig. 2 shows a measurement environment when investigated the brightness changes during horizontal viewing position changes... Is the distance of a typical display of the optimal viewing distance of 3 x display height called (H), taken with the presence of circular polarization filter at the position of the total of 12 points so as to be θ = ± 60 ◦ on 1600 mm.Figure 2: Experiment of the change in the brightness values for the horizontal angle θ vertical direction on which was 90 ◦ rotate the display in the survey is in the position of a total of six points so take the same position at θ = ± 30 ◦ and the method of FIG. 2 It was taken... At this time, using a laser distance meter MK-LAK for determining the distance and angle Co. Samurai Marketing Company... The angle of the calculation is a function of this device, we were using the theorem mode of capitalist Golas... Brightness value (luminosity) is to shoot an area of 25 cm 2 center of the display by a camera to obtain a luminance value of the pixel by adding the RGB color values of each pixel of the image... Brightness value used for the data using the average value divided by parametric by adding the pixel intensity value of the entire region of interest... I was using a Toshiba Regza42Z8 is Henhikari-shiki 3D display in this experiment... The camera is a photographic equipment I was using NikonD750... As a preliminary experiment, try the shutter speed by {1 / 500, / 250 / 125, 1 / 60, 1 / 30, 1 / 15, 1 / 8, 1 / 4, 1 / 2}, each shutter speed and The difference between the I and the reference value 1 / 30 that fall within a certain range... From preliminary experiments, image had became Cho properly black and white in the 1 / 250 and the outside of value than 1 / 8... F value at the time of shooting f/16, ISO was unified at 400... And shooting pattern that does not pinch the circular polarization filter between the camera and the display, we've taken two patterns of shooting pattern sandwiching the circular polarization filter... 3.2 are valued spread 4 K resolution support Disupu example is even now display market of display characteristics, 3D has been adapted to be integrated into the jar good for granted as part of the function, the ever-accelerating this momentum in the probably the future. In this section we will use the model shown from consumer products of 3D display that currently enter the hand in Table 1, as the center of the setting of brightness and contrast, and describe the results of the investigation the appropriate settings for the vision hearing of ExPixel content. Table 1: the target is the 3D display type name No... type the display name1 Toshiba REGZA42Z8 2 SONY BRAVIA KDL-42 W800B 3 LG Smart TV 42LB6700 to return the first each product in Table 1 in the state and me was a "calibration before" state, this survey of the setting of the factory state... If the first "non-use of circularly polarized light filter (no filter) in each display of calibration before state, were investigated in 2 pattern in the case of"yen use a polarizing filter (the filter there)... Then, the calibration of each display is performed using the datacolor's Spyder4Elite, all of the display was a "calibration" state... Similarly, if 'a' ¥ non use polarizing filters "in each display performs investigation in two patterns in the case of"yen using the polarizing filter"to compare the brightness value of each state... 4 were examined for changes in brightness value due to the distance between the experimental display and an observer of the specific 4.1.1 distance it has on the brightness value impact of the results 4.1 optimal viewing area. Normally the display luminance determined by ISO, in order to utilize the results obtained in this experiment with a wide range of displays, and utilizing paired elephant pixel values from the image photographed by the single-lens camera as the Luminosity (level). The results of the distance and the display brightness between the display viewer in FIG. 1 of the environment it is shown in Figure 3. Figure 3: Distance-luminance (circle without polarizing filter) from FIG. 3, the display-in between from the inter-camera distance 1000 mm of 2400 mm, it is almost a straight line of the graph there is no rapid change... This contrast in relation to the normal brightness and illuminance are considered as a radiate of the point light source, the number of displays are popular, the directivity of the backlight is controlled, that behaves as a parallel light source shows... (Which is measured in that order such proposed method.) Figure 4 shows the relationship between the intensity and the distance in the same condition when it is fitted with a polarizing filter to the camera lens. Figure 4: Distance-brightness (with circularly polarized light filter) in the case of digital filter, the overall brightness is down but I have not seen extreme decrease in brightness due to distance... Overall value, since it has fallen to half of FIG. 3, it found that the quantity of the thus cut-off pixel columns circularly polarizing filter fell. Although the phenomenon intensity fell, the brightness of the image photographed by the blocking of pixel columns corresponding to the circular polarization filters fell a (Fig. 5), a passive type circular widely used in a multiplexed invisible image polarizing filter Display and yet not become a factor to attenuate the intensity, even independent of the distance... Figure 5: circularly polarized light pixel column of the corresponding to the polarization filter is blocked I investigated the change in luminance due to the change in the horizontal position of the observer with respect to 4.1.2 Experiment display luminance value by the horizontal angle... Is the distance of a typical display of the optimal viewing distance of 3 x display height called (H), were taken at the position of a total of 12 points so get an angle ± 60 ◦ on 1600 mm (Figure 2)... Squarely to the display from FIG. 6, it is I say that lowered brightness by going angled... Figure 6: horizontal angle θ-brightness (without circularly polarized light filter) 7 distance even when it was passed through the circular polarization filter from the experiments and the like, but overall brightness has fallen, as the trend of the graph circularly polarized light filter is the same as when it is not passed... Figure 7: horizontal angle θ-brightness (with circularly polarized light filter) I investigated the changes in brightness due to 4.1.3 change of the viewer's vertical an angle θ to the experimental display of brightness value by the vertical angle... It is the same position at θ = ± 30 • the method of Figure 2 in terms of was 90 • rotating the display is in this study were taken at the position of the six points as. Similarly the horizontal direction in the vertical direction if the angled, reduction of the luminance values due to the angle get was confirmed (Fig. 8, Fig. 9)... Figure 8: (no circularly polarized light filter) vertical angle θ-brightness Figure 9: vertical angle θ-brightness (with circularly polarized light filter) for 4.2 compared to existing popular passive type 3D flat panel luminance value of each display, the default the luminance value of the video, the change in luminance value of the calibration settings to gamma = 2.2 has been targeted at ExPixel were compared... 10 when passing through the circular polarization filters before and after calibration, represents the value of the brightness value... Method of calculating the number of values is same as the method used for the experiment in the viewing range, taking the average of the values of each pixel intensity... Figure 10: changes in the brightness values before and after calibration display of each company from Figure 10 by default state, it can be set to high brightness value

ผลลัพธ์ (อังกฤษ) 2:

Suzuki Hisataka 1), Suzuki lily Aya 2), Tsuda Ryotaro 2), Vimuktalop Pinyapach3), Jaruthien Nine3), Shirai Akihiko 4)
1) Kanagawa Institute of Technology and Information Engineering (〒 243-0292 Atsugi, Kanagawa Prefecture Shimoogino 1030, Hisataka @ shirai.la) 2) Kanagawa Institute of Technology School of Information and Media Department 3) Chulalongkorn University, Department of Imaging and Printing Technology (254 Phyathai Road, Patumwan, Bangkok Thailand. 10330) 4) Kanagawa Institute of Technology School of Information and Media Department of Associate Professor Summary: SUCH as A home-use Game machine, by the FPGA hardware implementation That CAN multiplex A Variety of HDMI input, Towards the Generalization of multiplexing Invisible video Technology the Possibility of good Ri wide range of Applications HAS spread " ExPixel ", A Popular type display Device We Went viewing Characteristics in the Chair and its Evaluation. The Distance Between the Viewer's display and, for changes in the display Luminance Due to the angle Against the display, an Experiment WAS Performed in case of USING the circularly POLARIZING filter. Also HAS Been Released from the display manufacturers, the video Settings of the existing passive type 3D display, and report the results of the Experiments Were Conducted for the Change in display Luminance When it IS calibrated to A default setting and gamma = . 2.2 Keywords: multiplexing, polarization, 3D display . while 3D display for home HAS Been generalized to 1. Introduction, the supply of contents CAN successfully Express A stereoscopic effect IS at Present, not kept PACE with the supply of the hardware When Ru look back the history, 3D stereoscopic image Tends to BE Sedated in there IS A Tendency to call A transient Boom 2-4 years [1]. On behalf of the 3D di splay the TV Market, Considering That the Increase in Products to Sell 4K Resolution , CAN not Obtain A certified Mezaru When Also this Boom IS Inclined calming. Against this background, we as A Technique for giving A new value Added to the Consumer of the 3D display, Proposes multiplexing Invisible image Technology "ExPixel" [2]. This Technique Utilizes the circular polarization and the even odd pixel columns of the Consumer passive type 3D flat-panel, it IS A Technique That CAN Completely Present Another image Against the circularly POLARIZING filter Wearer and naked-eye Viewer. Although Achieves multiplexing Invisible image as shader Technology That Operates in GPU-based initial Study, [3] By Also succeeded in FPGA hardware Reduction in recent years, it Enables multiplexing of existing Consumer Content It WAS. For Investigation of Appropriate viewing area on Which to Watch multiplexing Invisible image, good viewing of multiplexing Invisible image That put to A Variety of 3D display in Consumer Products 3D display in this report, Pervasive passive type 3D display and it IS to report the results of Experiments on the viewing Characteristics. Hisataka SUZUKI, Yuria SUZUKI, Ryotaro TSUDA, Vimuktalop Pinyapach, Jaruthien Nine, and Akihiko SHIRAI 2. related Research multiplexing the video Technology HAS Been attempted by many Researchers. By Mistry "ThirdEye" [4] CRT, it WAS A proposed multi-Scenario Screenings USING glasses for opening and closing the shutter in Synchronization with A screen SUCH as LCD projectors at 120 Hz or Faster. It HAS Been already proposed by SUCH "Snared Illumination" by the high-speed shutter How by the division When USING the Fakespace Labs Inc. McDowall et al. [5] and multi Use Light Engine (MULE) projector. It IS not Carried concealed and Will BE Seen to overlap the Two Images to the naked eye of the Viewer IS "ThirdEye" but the Projection number Two movies image content in the "Snared Illumination" and by Providing A SUFFICIENT difference, A We Have REALIZED an Invisible video Technology in the Active shutter system. Although multi-player Game Technology USING THESE same Active shutter HAS Been Patented as "SimulView" than 2012 US Sony Computer Entertainment Inc. Also here both channels Are multiplexed multiple display [6]. The The moreover, the Characteristic Investigation of ordinary liquid crystal flat panel Displays Have Been Reported in 2007 by Yamamoto et al. [7]. On the Other hand this Mountain Are Reported Also for Characteristics and image quality Evaluation Technology of flat -screen display in [8]. 3. Experiment First ExPixel Investigation of the optimal viewing area for viewing multiplexing content with, and in A Variety of Consumer Products 3D display, in order to Obtain an optimal display of the ExPixel content video Settings Were Evaluated. 3.1 Were Subjected to the Following in three Experiments to Investigate the Investigation optimal viewing area of the Optimum viewing area. In the passive-type 3D display device various required by ExPixel, Along with the Presence or absence of A filter, Experiment 1 ) Disupu Ray and the observer's distance of (L), Experiment 2) horizontal position of the observation person with respect to the display (angle θ), Experiment 3) the vertical position of the observation person for display (the angle θ), to measure the brightness of the display When it IS Viewed at Each position for the three Experiments. Measurement position on the di splay copying in [8], and the Center of the screen Certain Areas. In this Experiment IS Masked in the display Center, an area in 25 cm 2, Which WAS USED as A Representative value of the Luminance of the Captured display by A digital single-lens NikonD750. By the reflection of the position and Displays, experimental environment for error Due to the natural light IS Generated WAS A darkroom. The camera IS Installation on A tripod to display and Capture an image at Each position Wants to measure the brightness. Figure 1 Shows A Measurement environment of the linear Distance, in 200 mm Intervals Between 2400 mm from 1000 mm to Orthogonally Against the Center . of the display window, and Were taken in all eight Locations Figure 1: Distance confirm the Change in the Luminance value Due to .. the Experiment 2, Fig 2 Shows A Measurement environment When Investigated the brightness changes During horizontal viewing position changes Is the Distance of A typical display of the optimal viewing Distance of 3 × display height called (H), taken with the Presence of circular polarization filter at the position of the Total of 12 points so as to BE theta = ± 60◦ on 1600 mm. figure 2: Experiment of the Change in the brightness values ​​for the horizontal angle theta vertical direction on Which WAS 90◦ rotate the display in the survey IS in the position of A Total of six points so Take the same position at theta = ± 30◦ and the method of FIG-. 2 It WAS taken. At this time, USING A Laser Distance Meter MK-LAK for Determining the Distance and angle Co. Samurai Marketing Company. The angle of the Calculation IS A function of this device, Were we USING the theorem mode of capitalist Golas. Brightness value (luminosity) IS to Shoot an area of 25 cm 2 Center of the display by A camera to Obtain A Luminance value of the pixel by Adding the RGB color values ​​of Each pixel of the image. Brightness value USED for the data USING the average value Divided by parametric by Adding the pixel Intensity value of the Entire region of interest. I WAS USING A Toshiba Regza42Z8 IS Henhikari-shiki 3D display in this Experiment. The camera IS A Photographic equipment I WAS USING NikonD750. As A preliminary Experiment, try the shutter speed by {1/500, / 250, / 125, 1 / 60,1 / 30,1 / 15,1 / 8,1 / 4,1 / 2}, Each shutter speed The difference Between the I and the reference value 1/30 That fall within A Certain range. From preliminary Experiments, image HAD Became Cho Properly black and white in the 1/250 and the outside of value than 1 / 8. F value at the time of Shooting F / 16, ISO WAS Unified at 400. And Shooting pattern That does not pinch the circular polarization filter Between the camera and the display, we've taken Two Patterns of Shooting pattern Sandwiching the circular polarization filter. 3.2 Are valued spread 4K Resolution Support Disupu example IS even Now display Market of display Characteristics, 3D HAS Been Adapted to BE integrated into the JAR good for granted as part of the function, the ever-Accelerating this momentum in the Future probably. In this section we Will use the model shown See from Consumer Products of 3D display That currently enter the hand in Table 1, as the Center of the setting of brightness and contrast, and describe the results of the Investigation the Appropriate Settings for the Vision Hearing . of ExPixel content Table 1: target IS the 3D display type name No. display type name 1 Toshiba REGZA42Z8 2 SONY BRAVIA KDL-42W800B 3 LG Smart TV 42LB6700 to Return the first Each Product in Table 1 in this survey of the setting of the Factory state, the state me WAS A "calibration before" state. If the first "non-use of circularly polarized light filter (no filter)" in Each display of calibration before state, Were Investigated in 2 pattern in the case of "Yen use A POLARIZING filter (filter there) ". Then, the calibration of Each display IS Performed USING the Datacolor's Spyder4Elite, all of the display WAS A "calibration" state. Similarly, if 'A' ¥ non use POLARIZING Filters "in Each display Performs Investigation in Two Patterns in the case of "Yen USING POLARIZING filter" to compare the brightness value of Each state. 4. Were Examined for changes in brightness value Due to the Distance Between the experimental display and an observer of the specific 4.1.1 Distance it HAS on the brightness value impact of the results 4.1 optimal viewing area. Normally the display Luminance Determined by ISO, in order to Utilize the results Obtained in this Experiment with A wide range of Displays, and Utilizing paired elephant pixel values ​​from . the image Photographed by the single-lens camera as the Luminosity (level) . The results of the Distance and the display brightness Between the display Viewer in FIG-1-of the environment it IS shown See See in Figure 3. Figure 3: Distance - Luminance (Circle without POLARIZING filter) from FIG--3, the display -.. in Between from the inter-camera Distance 1000 mm of 2400 mm, it IS Almost A Straight line of the Graph there IS no rapid Change This contrast in relation to the normal brightness and Illuminance Are Considered as A Radiate of the point light source, the number of Displays Are Popular, the Directivity of the Backlight IS controlled, That Behaves as A Parallel light source Shows. (Which IS Measured in That order SUCH proposed method ..) Figure 4 Shows the relationship Between the Intensity and the Distance in the same condition When it IS Fitted with A POLARIZING filter to the camera lens Figure 4: Distance - brightness (with circularly polarized light filter) in the case of digital filter, the overall brightness IS down but I Have not Seen Extreme Decrease in brightness Due to Distance. Overall value, since it HAS fallen to Half of FIG-. 3, it found That the Quantity of the THUS Cut-off pixel columns circularly POLARIZING filter fell . Although the Phenomenon Intensity fell, the brightness of the image Photographed by the blocking of pixel columns CORRESPONDING to the circular polarization Filters fell A (Fig 5.), A passive type circular Widely USED in A multiplexed Invisible image POLARIZING filter Display and Yet not Become A factor to Attenuate the Intensity, even independent of the Distance. Figure 5: circularly polarized light pixel column of the CORRESPONDING to the polarization filter IS blocked I Investigated the Change in Luminance Due to the Change in the horizontal position of the observer with Respect to 4.1.2 Experiment display Luminance value by the horizontal angle. Is the Distance of A typical display of the optimal viewing Distance of 3 × display height called (H), Were taken at the position of A Total of 12 points so Get an angle ± 60◦ on 1600 mm (Figure 2). squarely to the display from FIG-6, it IS I say That Lowered brightness by going angled.. Figure 6: horizontal angle theta brightness (circularly polarized light without filter) 7 Distance even When it WAS passed through the circular polarization filter from the Experiments and the like, but overall brightness HAS fallen, as the trend of the Graph circularly polarized light filter IS the same as When it IS not passed. Figure 7: horizontal angle theta brightness (with circularly polarized light filter) I Investigated the changes in brightness Due to 4.1.3 Change of the Viewer's vertical angle theta to the experimental display of brightness value by the vertical angle. It IS the same position at theta = ± 30◦ the method of Figure 2 in terms of WAS 90◦ Rotating the display IS in this Study Were taken at the. position of the six points as Similarly the horizontal direction in the vertical direction if the angled, Reduction of the Luminance values ​​Due to the angle Get WAS Confirmed (Fig 8, Fig 9 ..). Figure 8: (no circularly polarized light filter) vertical angle theta brightness Figure 9: vertical angle theta brightness (with circularly polarized light filter) for 4.2 Compared to existing Popular passive type 3D flat panel Luminance value of Each display, the default the Luminance value of the video, the Change in Luminance value of the calibration Settings to gamma = 2.2 HAS Been targeted at ExPixel Were Compared. 10 When Passing through the circular polarization Filters before and after calibration, Represents . the value of the brightness value Method of Calculating the number of values ​​IS same as the method USED for the Experiment in the viewing range, taking the average of the values ​​of Each pixel Intensity. Figure 10: changes in the brightness values before and after calibration display of Each Company from Figure 10 by default state, you CAN BE set to high brightness value

ผลลัพธ์ (อังกฤษ) 3:

Yuri Suzuki Aya Suzuki Kuki (1) and (2), Tsuda (2) and (3) and Vimuktalop Ryotaro Pinyapach Jaruthien Nine Shirai Akihiko)
(3) and 4 (1), Graduate School of engineering, Kanagawa Institute of technology information (243 Atsugi City, Kanagawa Prefecture. 0292 outcrop 1030, Hisataka Shirai, La. (2) the information media, Kanagawa Institute of technology, Faculty of Chulalongkorn (3) University, Department of Imaging and Printing Technology (254 Phyathai Road, Patumwan, Bangkok Thailand.10330 (4), associate professor, Department of computer science, Faculty of information, Kanagawa Institute of technology, media is the outline of a game machine for home ， ， by FPGA hardware can be multiplexed HDMI input for various ， generalization "invisible" ExPixel multiplexing; you may spread to a wide range of applications evaluation about the characteristics and viewing in diffusion type display and display the distance and the mole ， viewers on the display by the angle with respect to the change of luminance of the display ，This experiment was carried out in the case of using circularly polarized light filter and ， released from the manufacturer of the display ， of existing passive type 3 D video display is set to the default settings and calibration of gamma 2.2 reported results of experiment about the brightness of display keyword: change of polarization multiplexing ， ， D display 3 is the introduction of the household, but generally D display ，Which is made of solid content of supply at present, catch up with supply of hardware is not from the history and ， 3 D stereoscopic video, best TV markets tend to call the transient boom tends to two to four years to calm on behalf D Di splay ， ， 4 resolution K think more products to sell in the current boom, authentication, and leaning on sedation; have no choice but to such an ， from backgroundWe ， new D display of consumer products and technologies to give added value to the ， using the even / odd pixel column and circular polarization in consumer products D passive type flat panel 3 in [2] n. this technology is proposed ExPixel "invisible" multiplexing the technology which can be presented to the naked eye image of a completely different viewers and circularly polarized light filter is installed at the beginning of the study ， operates in GPU base which realizes multiplexing of the invisible image as a shader TechnologyIt was so successful in the FPGA hardware can be multiplexed per existing con Schumer and content in recent 3 D appropriate consumer products on display to the multiplexed in the invisible image region in this paper, we study ， for the good of the invisible watch video multiplexing ， in D display various viewing characteristics of three popular type passivation ブ 3 D display experiment result is reported to the.
Hisataka SUZUKI, Yuria SUZUKI,Ryotaro TSUDA, Vimuktalop Pinyapach, Jaruthien Nine, and Akihiko is SHIRAI, many researchers have been attempted by the multiplexing technology research association "ThirdEye creatively by Mistry CRT ，LCD project and using glasses in synchronism with the screen to the opening and closing of the higher ‐ speed of more than Hz multiple scenarios proposed screening was carried out at this time by using a fast segmentation method with shutters Fakespace Labs company in McDowall and Snared Illumination [5]. Multi Use Light Engine MULE hiding some of ThirdEye "has been proposed by the naked eye of the viewer in the projector, video overlap of two looks， "Snared Illumination" two sufficient number of projected image content difference have realized the invisible image in a more shutter system ， [6] n. in the ルチプレイヤーゲーム technology and active shutter using the same Sony USA 2012 Computer Entertainment "company" as SimulView obtained patents and two channel multiplexing multiple displaysThe conventional LCD flat panel display characteristics in [7] n. on the other hand, the survey by Yamamoto et al. Reported in 2007 on the characteristics and quality of flat panel display technology, the mountains have report of 8].



optimal viewing area for viewing the multiplexing experimental investigation on content ， ExPixel first and a variety of consumer goods in 3 D display image is set to ， evaluation was carried out in order to obtain the optimal ExPixel display content. 3.An experiment was carried out to study the optimal viewing area of research for optimal viewing area in ， passive type display device 3 D ExPixel required in various experiments and filter with ， snipe displ (1) and the distance (2) and L experiment system study on the display position and a horizontal angle. Theta. (3) of experimental study on the display position of the vertical angle. Theta ，In three experiments in view of the position of measuring the luminance of the display 8 in the measurement position on the day following the ， spray to the center of the screen in a predetermined region in the present experiment in the central display masks ， 25 cm in ， region in NikonD 750 digital single lens is used as a representative value of the brightness of the display ， by reflection position, and play dayDarkroom and actual test environment for the error caused by natural light in the display device provided with a camera on a tripod ， image was taken at each position to measure the brightness ， and distance measurement environment of Fig. 1 is perpendicular to the center of the display window as 1000 mm from 24 mm to 200 mm apart from ，.
figure 1 was performed in all eight points in the distance of the luminance value changes in Experiment 2 ，Fig. 2 is a survey of the measuring environment change shows the bright degree of the change of the position of the horizontal direction are commonly referred to as the optimal viewing distance of the display is the display height. H ， distance.
Figure 2 in the presence or absence of a circular polarizing filter a total of 12 point so that. Theta. = 60 ◦ 1600 mm on the experiment at the vertical direction change of luminance values for horizontal angle of a display which is rotated by 90 ◦ atAs shown in Fig. 2 and the way that ◦. Theta. 30 at the same position in a total of six points with Samurai Marketing Company Ltd. MK LAK laser distance meter and to determine the distance and the angle at the time of ， angle calculation of the function device using the theorem of pita Athenagoras mode the luminosity luminance value of the central ， display area 25 cm 2 photograph ，RGB of each color of each pixel in the image and the luminance value is added to the value of the pixel data based on the luminance value is added to the whole area of the pixel brightness value of a parameter object divided by the average value is used in the present experiment is D polarization equation (3) the display 42 Toshiba Regza Z 8 with the camera equipment to be used NikonD 750. ， as preliminary experiments of the shutter speed is 1 / 500 ， 1 / 1 / 125, 250 ， ， 1 / 1 / 30 / 60 ， ， 1 1 / 8 / 15 ， ， 1 / 4 ， 1 / 2 test ，On the basis of the values within the range of 1 to 30, the difference between the shutter speed by a preliminary experiment ， 1 1 / 8 / 250 or more outside the image has become a publisher or black and white as the value F F 16 ， photographing the ISO unified pattern between 400 and circular polarization filter between the camera and the display ， shot shot pattern of circularly polarized filter 2. 3.Evaluation of display characteristics in the current market K display resolution spread corresponding snipe displ ray and ， ， but it is of course as part of it's function is incorporated in the three D this momentum will follow the way to use from the acceleration model D display of consumer products are shown in Table 1, 3 is now available in this section and ， brightness and trust it's set ， for visual auditory ExPixel content will be set to the survey results of.
Table 1At the No type 3 D display for the display name of the Z Toshiba REGZA 42 8 1 2 SONY BRAVIA KDL 42 W 800 B 3 LG Smart TV 42 LB 6700 survey at first ， returns in the factory to the product of its set state "before calibration and calibration of each display in the" first state "before using circularly polarized filter (filter)" ，The "circular polarization filter (filter) was carried out in two patterns of the" If then ， of each display キャリブレーショ Datacolor's using Elite Spyder 4 ， all display "calibration" state in the same way as in each of the circularly polarized light to display "" no use "filter" ， survey was carried out in a circular pattern is used to filter the ， like 4. Compared with the luminance value of each state results. 4.A certain distance of the optimal viewing area 4.1.1 by distance experiment display luminance value and effect observation of the change of luminance values with ， display luminance is usually obtained by ISO for use in a wide range of this play ， results obtained in this experimentLuminosity values for pixels from the image photographed by the camera level used as in Fig. 1 display environment among viewers and the distance of the bright display results are shown in Fig. 3 in Fig. 3.
distance from the luminance ， circularly polarized HTS filter (NO 3 from the distance between the camera and the display ， 1000 between 2400 mm without abrupt changes in the ， has become almost a straight line in this graph ， is considered as the luminance and illuminance of the light source and to releaseMany of the popular ， ， directivity of the back light is controlled so as to.
polarization filter in Fig. 4 shows the relationship of distance and in the same condition when it is mounted to the camera lens of Fig. 4 measured by hand method behaves parallel light source and shows that using this distance (circularly polarized HTS filter, filter, there is a case that the luminance of the entire ， extreme brightness measurement from falling but not in the overall value ，Fig. 3 ， falls in half, it is found that the amount of the fallen column of pixels by circularly polarized light filter which fell into the luminance and luminance phenomena for ， fall on an image photographed by a block of pixels corresponding to the circular polarizing filter Fig. 5 is ， multiplexing is widely used in the invisible image at the passive type circular polarizing filter to attenuate the intensity factors of the display is not ，.
5 -- it is not affected by the distanceThe workers look for the luminance value of the experimental investigation of display brightness according to the change of the horizontal position of the changes 4.1.2 horizontal angle corresponding to the circular polarization filter is blocked by the pixel row of general and optimal viewing distance of the display unit 3. Display height is the distance H. ， 1600 to 60 ◦ angle on a total of 12 mm at point ， squarely to the display in Fig. 6 (Fig. 2).Fig. 6.
goes down the brightness at that angle horizontal angle. Theta. (the circularly polarized circular graph of distance experiment has been ， down your entire ， brightness when the polarizing filter without filter, including through the circle from Fig. at the same time, and is impervious to the polarizing filter.
Figure 7 horizontal angle. Theta (including circular polarization filter is 4.1.We investigated the change of the vertical angle observation on the change of luminance values on the display by the vertical angle experiment in this study to display 90 to ◦ times as in the same way as in Fig. 2 and ◦ position. Theta. 30. With a total of 6 point position in the vertical direction of the field angle in the horizontal direction as well as in ， by angle can decrease the luminance values (Fig. 8 and Fig. 9 ， ）.
figure 8 vertical angle. Theta (Fig. 9 without circularly polarized HTS filter.The vertical angle. Theta. (there is a circularly polarized HTS filter 4.2 display luminance value of each of the existing popular type of passive type 3 D flat panel on ， de fault image and the luminance value of the target in the ， ExPixel and 2.2 gamma calibration set value of the change of luminance Fig. 10 is compared with that of before and after calibration by circularly polarized light filters through the ， luminance value of the value that represents the value of the calculation range of the viewing time and test method using the same ，Fig. 10.
by taking the average luminance value of each pixel of the luminance value before and after calibration changes from Fig. 10 at the display of each company ， default state is set to a higher value.