TW200908727A - Method and apparatus for dynamic contrast enhancement - Google Patents
Method and apparatus for dynamic contrast enhancement Download PDFInfo
- Publication number
- TW200908727A TW200908727A TW96129495A TW96129495A TW200908727A TW 200908727 A TW200908727 A TW 200908727A TW 96129495 A TW96129495 A TW 96129495A TW 96129495 A TW96129495 A TW 96129495A TW 200908727 A TW200908727 A TW 200908727A
- Authority
- TW
- Taiwan
- Prior art keywords
- histogram
- box
- contrast enhancement
- dynamic contrast
- preset
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000002156 mixing Methods 0.000 claims abstract description 48
- 241000023320 Luma <angiosperm> Species 0.000 claims abstract 3
- OSWPMRLSEDHDFF-UHFFFAOYSA-N methyl salicylate Chemical compound COC(=O)C1=CC=CC=C1O OSWPMRLSEDHDFF-UHFFFAOYSA-N 0.000 claims abstract 3
- 238000006243 chemical reaction Methods 0.000 claims description 35
- 230000007704 transition Effects 0.000 claims description 17
- 238000001914 filtration Methods 0.000 claims description 13
- 230000008569 process Effects 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 7
- 239000000203 mixture Substances 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000005728 strengthening Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 description 40
- 238000010586 diagram Methods 0.000 description 10
- 238000012546 transfer Methods 0.000 description 5
- 230000008859 change Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 238000010187 selection method Methods 0.000 description 2
- 230000016776 visual perception Effects 0.000 description 2
- 206010033557 Palpitations Diseases 0.000 description 1
- 230000004456 color vision Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 235000011389 fruit/vegetable juice Nutrition 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000010977 jade Substances 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Landscapes
- Studio Devices (AREA)
- Picture Signal Circuits (AREA)
- Image Processing (AREA)
Abstract
Description
200908727 九、發明說明: 【發明所屬之技術領域】 本1¾明有關於動態對比強化(dynamic contrast enhancement ),尤有關於一種混合一個以影像内容為主的 直方圖(content-based histogram )與一使用者自訂 (user-defined )直方圖之動態對比強化裝置及其方法。 【先前技術】200908727 IX. INSTRUCTIONS: [Technical field of invention] This is about dynamic contrast enhancement, especially about mixing a content-based histogram and a use. User-defined histogram dynamic contrast enhancement device and method thereof. [Prior Art]
以下為方便說明,在本說明書與申請專利範圍中,「亮 度」(luminance)表示—輸入影像資料值之一種成份 (component)’該成份與一顯示(displayed)影像資料值的 使用者肉眼感知強度(perceived intensity)有關,其中,該 輸入影像資料值本身可以是數位或類比之型式,其最終會 被員示在數位或類比顯示器上,例如陰極射線管(crt) 或液日日顯示(LCD)。「色度」(chrominance)表示該輸入 ’v像貝料值之另一種成份,該成份與該顯示影像資料值的 使用者肉眼感知色彩(eQ㈣有關,至於,色調(hue)與飽和 | ati〇n)疋一種常用來界定色彩感知(perception)的 標準’二者共同定義了色度。另外,「動態對比強化」表 八柜據〜像冗度分布,將過暗或過亮的影加以能 整?「使用者自訂直方圖」表示已預先被定義, 或在貫施過程中可被動態程式化的直方圖。 由於轉換函數(tmnsfer functi〇n)必需動態地根據影像内 :之:旁其調整之複雜程度可想而知。以影像強化為目 、 处理領域中,許多有關於影像對比調整的方法彳目 繼被研菸Ψ a 〗υ正α乃忐相 Χ來,最著名的應該是直方圖等化法(hiSt〇gram 200908727 equalization)。現存技術多根據直方圖等化法而改良,大致 可刀為兩類.曲線選擇方法與曲線運算方法。曲線選擇方 法是預先提供一組轉換函數,再根據動態影像内容從該組轉 換函數中選取其中之一,此種技術以美國第5, 8〇8, 號 專利文獻為代表。然而,影像内容組成相當複雜,恐難由 預。又的組轉換函數中擇其一而得到正確補償,故實際應用 - 範圍有所限制。 "曲線運算方法雖不需預先提供一組轉換函數,而係根據 c 純粹動態運算而得到—轉換函數,但此種方法多有調整過劇 以致於壳度均值(mean value)偏移過大的缺點,例如,在 夜晚擷取的影像有如白天。此外,該曲線運算方法還會使 得π度位階(level)之漸層變化(gradient changes)不連 續,此類技術以 Kim 於 1997 年(,’Contrast Enhancement Usmg Brightness Preserving Bi-HistogramThe following is a convenient description. In the scope of this specification and the patent application, "luminance" means - a component of the input image data value - the user's visual perception intensity of the component and a displayed image data value. (perceived intensity), wherein the input image data value itself may be in the form of a digit or analog, which will eventually be displayed on a digital or analog display, such as a cathode ray tube (crt) or a liquid day display (LCD). . "chrominance" means that the input 'v is another component of the shell value, which is related to the user's visual perception color (eQ(4)) of the displayed image data value. As for the hue and saturation|ati〇 n) A standard commonly used to define color perceptions, which together define chromaticity. In addition, the "Dynamic Contrast Enhancement" table 8 cabinets according to the redundancy distribution, can be too dark or too bright to be able to rectify? A "user-defined histogram" means a histogram that has been previously defined or dynamically programmed during the process. Since the conversion function (tmnsfer functi〇n) must be dynamically based on the inside of the image: the complexity of its adjustment can be imagined. In the field of image enhancement, in the field of processing, many methods for image contrast adjustment have been studied. The υ α α 忐 忐 , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , 200908727 equalization). The existing technology is mostly improved according to the histogram equalization method, and the approximate knife can be two types. The curve selection method and the curve calculation method. The curve selection method is to provide a set of conversion functions in advance, and then select one of the group of conversion functions according to the motion picture content. This technique is represented by the US Patent No. 5, 8-8. However, the composition of the video content is quite complicated and it is difficult to predict. The other group conversion function is selected to be correctly compensated, so the actual application - the scope is limited. "The curve operation method does not need to provide a set of conversion functions in advance, but the conversion function is obtained according to the pure dynamic operation of c, but this method has been adjusted so that the mean value of the mean value is too large. Disadvantages, for example, are images taken at night like daytime. In addition, the curve calculation method also makes the gradient changes of the π degree level discontinuous. This technique is based on Kim in 1997 (, 'Contrast Enhancement Usmg Brightness Preserving Bi-Histogram
Equalization»,Y.-T. Kim,IEEE Transections on Consumer Electronics Vol_43, N〇1 Feb 1997)所提出的方法為代 表而’ K i in所提出的方法有兩個問題:第一個問題是 - 不同的影像亮度分布竟然有相同的亮度均值;第二個問題 . 是統計上,亮度均值對離群值(outlier)極其敏感,換言 之,相對較少數像素之亮度值雖都遠離亮度均值,卻對該 亮度均值有报大的影響。為解決上述問題,故提出本發明。 【發明内容】 有鑒於上述問題,本發明目的之一為提供一種動態對 比強化裝置,係混合一個以影像内容為主的直方圖與一使 6 200908727 用者自訂直方圖。 為達成上述目的,本發明之動態對比強化裝置包含: 一直方圖計算器,用以計算一目前頁框(frame)之—亮度直 方圖,低通渡波器U〇w-pass filter),連接至該直方圖叶 算器,用以對該亮度直方圖執行一低通濾波處理,並產生 一濾波直方圖;一混合單元,連接至該低通濾波器,根據 混合權值(blending weight ),以混合該濾波直方圖與 一使用者自訂直方圖,並產生一混合直方圖;以及,一亮 度映射器(remapper),連接至該混合單元,用來根據該混 合直方圖以產生一轉換函數,並根據該轉換函數以映射該 目前頁框。Equalization», Y.-T. Kim, IEEE Transections on Consumer Electronics Vol_43, N〇1 Feb 1997) The proposed method is representative and the method proposed by 'K i in has two problems: the first problem is - different The image brightness distribution has the same average brightness; the second problem is that the brightness mean is extremely sensitive to outliers, in other words, the brightness values of relatively few pixels are far from the brightness mean, but The average brightness value has a large impact. In order to solve the above problems, the present invention has been proposed. SUMMARY OF THE INVENTION In view of the above problems, it is an object of the present invention to provide a dynamic contrast enhancement apparatus that mixes a histogram based on image content and a user-defined histogram of 6 200908727. To achieve the above object, the dynamic contrast enhancement apparatus of the present invention comprises: a histogram calculator for calculating a current frame-brightness histogram, U〇w-pass filter, connected to The histogram controller is configured to perform a low pass filtering process on the luminance histogram and generate a filtered histogram; a mixing unit connected to the low pass filter according to a blending weight Mixing the filtered histogram with a user-defined histogram and generating a mixed histogram; and a remapper coupled to the mixing unit for generating a conversion function based on the mixed histogram, And according to the conversion function to map the current page frame.
本發明另一個目的為提供一種動態對比強化方法,包 含以下步驟:計算一目前頁框之一亮度直方圖;對該亮度 直方圖執行一低通濾波處理,以產生一濾波直方圖;根據 此&權值,混合§玄濾、波直方圖與一使用者自訂直方圖以 產生一混合直方圖;根據該混合直方圖以產生一轉換函 數;以及,根據該轉換函數以映射該目前頁框。 兹配合下列圖示、實施例之詳細說明及申請專利範 圍’將上述及本發明之其他目的與優點詳述於後。 【實施方式】 本發明之動態對比強化裝置可以利用硬體、軟體 '韌 體(firmware)之其中之一、或前三者之任意組合來實施, 例如:純硬體實施的例子為一現場可程式邏輯閘陣列 (field programmable gate array,FPGA)設計、或一 特殊應用積體電路(application specific integrated 7 200908727 circuit ’ ASIC)設計,而硬體與韌體合併實施的例子為一 數位jg號處理器(digital signal processor,DSP)及其内建 韌體之組合。 第1圖為本發明第一實施例的架構示意圖。參考第1 圖’本發明動態對比強化裝置100包含一直方圖計算器 110、一低通濾波器12〇、一權值計算器13〇、一混合單元 140、一正規器(norma lizer) 170以及一亮度映射器150。 因為一目前頁框160是由多個輸入影像資料(或像素)所Another object of the present invention is to provide a dynamic contrast enhancement method comprising the steps of: calculating a luminance histogram of a current page frame; performing a low pass filtering process on the luminance histogram to generate a filtered histogram; Weight, mixed § 滤 filter, wave histogram and a user-defined histogram to generate a mixed histogram; according to the mixed histogram to generate a conversion function; and, according to the conversion function to map the current page frame . The above and other objects and advantages of the present invention will be described in detail below with reference to the accompanying drawings. [Embodiment] The dynamic contrast strengthening device of the present invention can be implemented by using any one of hardware, software 'firmware, or any combination of the first three. For example, an example of a pure hardware implementation is a field. A field programmable gate array (FPGA) design, or a specific application integrated circuit (application specific integrated 7 200908727 circuit ' ASIC) design, and the hardware and firmware implementation example is a digital jg processor (digital signal processor, DSP) and its built-in firmware combination. Figure 1 is a block diagram showing the architecture of a first embodiment of the present invention. Referring to FIG. 1 , the dynamic contrast enhancement apparatus 100 of the present invention includes a histogram calculator 110, a low pass filter 12A, a weight calculator 13A, a mixing unit 140, a norma lizer 170, and A brightness mapper 150. Because a current page frame 160 is composed of multiple input image data (or pixels)
組成’每一輸入影像資料值可區分為一亮度成份與一色度 成份,直方圖計算器丨丨〇在接收完目前頁框丨6〇的所有像 素之亮度成份(或資料)之後,計算出目前頁框16〇之一亮 度直方圖。 低通濾波器120接收該亮度直方圖後,即對該亮度直 方圖進行—低通濾波處理,以產生一濾波直方圖。第2圖 疋兩個轉換函數的比較圖,其中,轉換函數22是將亮度 直方圖21直接等化(equalized)而得,並未執行低通濾 波處理’·至於轉換函數24的產生是先將亮度直方圖Μ傳 运至低通;t波mx執行低通濾波處理,並得到遽波直 =圖23後’再將該渡波直方圖23等化而得。比較兩個轉 、函數22、24,可以明顯看到,轉換函數22的漸層變化 相阜乂於轉換函冑24的漸層變化,比較不連續。換言之, 存下來轉;Γ函數22,轉換函數24有較多的漸層層次被保 (稍U 在後續程序中正確決定—混合權值之大小 (稍後再討論)。 根據本發明,使用者自訂直方圖係已預先被定義,以 8 200908727 及/—或在實施過程中可被動態程式化。例如,使用者可以 為每-箱型區(bin)(直方圖係由多個箱型區所組成,每一 相型區均代表一個特定亮度範圍)指定一個強度 (magimude)或計數值,以預設—使用者自訂直方圓。為了 混合濾波直方圖與使用者自訂直方圖,此二者在進入權值 計算器130之前必須先被正規化(n〇rmalize)至一相同的 參考基準’例如:256 (亮度位階總數)。依此種方式,若 濾波直方圖的像素總數目多達bt〇tal,則正規器丨7〇所進 行的正規化處理就是將濾波直方圖從總數目bt〇tal等比 例縮小至256。舉例而言,若L (亮度位階總數)=256以 及bsize(總箱型區數目)=8時,則每一箱型區寬度(或灰 階級距)=32亮度位階。首先,正規器i 7〇計算濾波直方 圖的像素總數目btotal = |;—細⑺,其中bin⑴是濾波直方圖 第1個箱型區的計數值。接著,正規器17〇藉由計算bin(i) -(bin (i) X btotal)以正規化每—箱型區之計數值, 其中,i為一整數且涵蓋從0到(bsize-i)的範圍。請注 意’若該濾波直方圖與該使用者自訂直方圖在被送入權值 计异益130之前已具有相同參考基準,亦即具有相同的亮 度位階總數’則本發明動態對比強化褒置1 00中之正規哭 1 7 〇可以被省略。 權值計算器1 3 0接收濾波直方圖’並根據濾波直方 圖、使用者自訂直方圖與一預設混合參數办βΖ·„與比例之 關係曲線(稍後再討論),來產生一混合權值。混合單 元1 40再根據該混合權值sw ’以混合該渡波直方圖與該 使用者自訂直方圖’並產生一混合直方圖。第3圖係顯示 200908727 以產生一混合 混合一濾波直方圖與一使用者自訂直方圖 直方圖的—個例子。以T,請根據第3圖與第4A〜4C圖 說明權值計算器130與混合單元14〇之運作,其中,亮度 位階總數=2 5 6,箱细區她砉f - j4 a· 相玉[-㈣數-4,故母—箱型區寬度=64亮 度位階。 =參考第3圖,遽波直方圖31之任一箱型區中所擁 有的最大汁數值(以下稱之為)等於1Μ,而使用 者自。丁直方® 32之>[壬-箱型區中所擁有的最大計數值 (以下稱之為等於64,參數、卿所代表 的計數值大小皆由權值計算器13()所決定。 權值計算器1 3〇係根據三條混合參數办與比例之 關係曲線之其中之一及一比例值(=叩)之大 小,來產生一混合參數办αζ·„。本發明提供三條混合參數 與比例之關係曲線,如第4A〜4C圖所示,形狀分別是梯 形(trapezoid)、樓梯型(staircase)以及非線性(n〇n Hnear)。 三條混合參數與比例之關係曲線的趨勢大致上相似,亦 即,當比例值(=接近i時,表示濾波直方 圖與使用者自訂直方圖的亮度分布類似,混合參數 被設為1 ;而當比例值越來越大時,表示混合直方圖與濾 波直方圖的亮度分布越來越相似,混合參數的值也 隨之慢慢變大。最後,當該比例值大於一臨界值時,混合 參數办ah就被設為一常數(以第4A〜4C圖為例,該常數 等於2)。然而,在應用實施時,三條關係曲線分別具有不 同程度的困難度:樓梯型關係曲線(第4B圖)最容易被 實施’相對上’非線性關係曲線(第4C圖)實施的困難 10 200908727 度最南。假設混合參數办⑴·„ (=1 _5)之產生係依據一比 例值(=maxZ^/wep=128/64 = 2,如第3圖所示)與第4Λ〜4C 圖中三條混合參數與比例之關係曲線之其中之一。 根據上述參數,權值計算器13〇 即可動態產生一混合權值SW。在本實施例中,權值計算 器130係利用以下的數學方程式來決定混合權值: maxbin x in; + step x (1-^) = dgain x step...........(A) 權值§十鼻器13 0只需把28,以及 办d«=1.5代入方程式(A),就能得到混合權值聊=〇·5。接著, 混合單元140根據計算出的混合權值⑽(=〇 5),混合渡波 直方圖3 1與使用者自訂直方圖3 2,以一次混合一個箱型區 的方式(bin by bin)’來逐步建立或產生一個混合直方圖 33。依此方式,就可得到混合直方圖33的每一個箱型區之 最終計數值,例如:混合直方圖33的第〇個箱型區的最終計 數值等於96 (=128 X 0_5 + 64 X 〇·5),而第1個箱型區的 最終計數值等於64 (=64 X 0_5 + 64 X 0.5)等等。在方程式 (A)中,(_ζ>2Χ^φ)的功能係用來限制每一個箱型區的計數 值’以避免亮度分佈調整過劇與過度強化的情形產生。 最後’亮度映射器17 0接收該混合直方圖之後,執行 積分運算以產生一轉換函數,再根據該轉換函數以映射目 前頁框160。 需注思的是’本發明之參數w 、並非只倉匕分 別代表(不受限於)濾波直方圖與使用者自訂直方圖之任一 箱型區中所具有之最大計數值。所似心·^可以是選自一預設 乾圍内之任一值,該預設範圍是濾波直方圖之任一所有箱 200908727 型區中所具有之最大計數值至第N大計數值,其中,N 等於該濾波直方圖之箱型區總數除以3且N為一正整數。 舉例而言,如果該濾波直方圖之箱型區總數等於15,則 可以是選自濾波直方圖之任一箱型區中,所具有之 最大計數值至第5(N=15/3=5)大計數值的五個計數值中的任 一個計數值。此外,所似心·/?也可以是一預設範圍内之均 值’該預設範圍是濾波直方圖之任一箱型區中所具有之第 Μ大計數值至第N大計數值,其中,N等於該濾波直方圖 之箱型區總數除以3,Μ、N均為正整數且N >M。例如: 當M=2且該濾波直方圖之箱型區總數等於15時,則所似办以 疋濾波直方圖之所有箱型區中,所具有之第二()最 大计數值至第5(N==l5/3=5)大計數值的五個計數值間之一個 均值。同樣地,可以是一預設範圍内之任一值,該預 設範圍是使用者自訂直方圖之任一箱型區中所具有之最 大計數值至第P大計數值,其中,p等於該使用者自訂直 方圖之箱型區總數除以3且p為一正整數。另外,“印也 可以是一預設範圍内之均值,該預設範圍是使用者自訂直 方圖之所有箱型區中所具有之第Q大計數值至第P大計數 值,其中,P等於該使用者自訂直方圖之箱型區總數除以 3 ’ Q、P均為正整數且p >Q。 第5圖為本發明第二實施例的架構示意圖。參考第5 圖’本發明動態對比強化裝置5〇〇包含一直方圖計算器 110、一低通濾波器12〇、一混合單元14〇、一直方圖限制 器5〗〇、一時域(temp〇ral)濾波器520、一亮度映射器150、 正規器170以及一場景偵測器(scene detector)530。相 12 200908727 較於第一實施例,第二實施例的動態對比強化裝置5〇〇雖 少了權值計算器130,但多了直方圖限制器51〇、時域遽 波器520以及場景偵測器53〇。至於在第二實施例與第一 實施例中具有相同標號之元件,同樣具備有相同之功能’ 在此不予贅述。 在第二實施例中,混合單元14〇係利用一固定 ' (Static)混合權值來混合濾波直方圖與使用者自訂 • 直方圖,以一次混合一個箱型區的方式,來逐步建立一個混 ( 合直方圖。換言之,binb⑴=binf⑴x⑽+ Μηυ⑴χ (1-謂),其中,binb⑴,binf(i)與binu⑴分別是混合直方 圖、濾波直方圖與使用者自訂直方圖中第丨個箱型區的計數 值。本發明為避免直方圖亮度分佈調整過於劇烈,故設置 一個直方圖限制器510於混合單元14〇之後。直方圖限制 為5 1 0限制混合直方圖中每一箱型區的計數值必須落在一特 定範圍:(G1 X价户)與(G2 X wep)之間,其中,“印代 表使用者自訂直方圖之所有箱型區中所具有之最大計數 ;值,G1、G2均為正整數且G2 >G1。也就是說,若混合直 .方圖之任一箱型區的原始計數值超出(G2 χ “卬),則該 箱型區的最終計數值會被直方圖限制器5 1 〇減為(G2 χ WM);反之,若混合直方圖之任一箱型區的原始計數值低 於(G1 χ ,則該箱型區的最終計數值會被直方圖限 制器510增加為(G1 χ 尸)。 請繼續參考第5圖’場景偵測器5 3 〇計算前後二頁框 間的直方圖分佈差距,並比較該差距與一臨界值τ,當該 差距大於該臨界值Τ時,場景偵測器53〇判斷已產生一場 13 200908727 景轉換(scene change )(亦即在一視訊序列(video sequence) (The composition 'each input image data value can be divided into a luminance component and a chrominance component, and the histogram calculator 计算 calculates the current luminance component (or data) of all the pixels of the current page frame 丨6〇 One frame brightness histogram of page frame 16〇. After receiving the luminance histogram, the low pass filter 120 performs a low pass filtering process on the luminance histogram to generate a filtered histogram. Fig. 2 is a comparison diagram of two conversion functions, wherein the conversion function 22 is obtained by directly equalizing the luminance histogram 21, and the low-pass filtering processing is not performed'. As for the generation of the conversion function 24, The luminance histogram Μ is transmitted to the low pass; the t wave mx performs low-pass filtering processing, and obtains the chopping straight=Fig. 23 and then equalizes the crossing histogram 23. Comparing the two transitions, functions 22, 24, it can be clearly seen that the gradual change of the transfer function 22 is relatively discontinuous compared to the gradual change of the conversion function 24. In other words, the function is saved; the function 22, the conversion function 24 has more gradient levels to be guaranteed (slightly U is correctly determined in subsequent procedures - the size of the blending weights (discussed later). According to the invention, the user Custom histograms have been pre-defined to be dynamically stylized at 8 200908727 and / or during implementation. For example, the user can be a bin-box (the histogram is made up of multiple bins) The area is composed, each phase area represents a specific brightness range.) Specify a intensity (magimude) or count value to preset - user-defined rectangular circle. In order to mix the filtered histogram and the user-defined histogram, Both of these must be normalized (n〇rmalize) to an identical reference datum before entering the weight calculator 130. For example: 256 (total number of luminance levels). In this way, if the total number of pixels of the filtered histogram is As long as bt〇tal, the normalization process performed by the normalizer 就是7〇 is to reduce the filter histogram from the total number bt〇tal to 256. For example, if L (total number of brightness levels) = 256 and bsize (Number of total box types) =8, then each box type area width (or gray level distance) = 32 brightness level. First, the normalizer i 7〇 calculates the total number of pixels of the filtered histogram btotal = |; - fine (7), where bin(1) is the filtered histogram The count value of the first box type area. Then, the normalizer 17 calculates the count value of each box type by calculating bin(i) -(bin (i) X btotal), where i is one Integer and covers a range from 0 to (bsize-i). Please note that if the filtered histogram and the user-defined histogram have the same reference before being sent to the weighted benefit 130, that is, The same total number of brightness levels 'the normal contrast in the dynamic contrast enhancement device of the present invention can be omitted. The weight calculator 1 3 0 receives the filtered histogram' and according to the filtered histogram, the user-defined histogram The graph and a preset blending parameter do βΖ·„ versus scale (discussed later) to generate a blending weight. The blending unit 1 40 further mixes the ferrometer histogram according to the blending weight sw′ The user customizes the histogram 'and produces a mixed histogram. 3rd The system displays 200908727 to generate a hybrid hybrid-filter histogram and a user-defined histogram histogram. For T, please describe the weight calculator 130 and the mixing unit according to the 3rd and 4A-4C diagrams. 14〇 operation, in which the total number of brightness levels = 2 5 6, box fine area her 砉 f - j4 a · phase jade [- (four) number -4, so the mother - box type area width = 64 brightness level. = Refer to Figure 3 The maximum juice value (hereinafter referred to as) in any box type of the chopping histogram 31 is equal to 1Μ, and the user is from the Dingzhifang® 32>[壬-box type zone The maximum count value (hereinafter referred to as equal to 64, the size of the count value represented by the parameter, qing is determined by the weight calculator 13 (). The weight calculator 1 3 产生 is based on one of the three mixed parameters and the proportional relationship curve and a proportional value (= 叩) to generate a mixed parameter ζ ζ „. The present invention provides three mixing parameters and The relationship between the ratios, as shown in Figures 4A to 4C, is trapezoid, staircase, and nonlinear (n〇n Hnear). The trend of the relationship between the three mixed parameters and the ratio is roughly similar. , that is, when the scale value (= close to i, the filter histogram is similar to the brightness distribution of the user-defined histogram, the blending parameter is set to 1; and when the scale value is larger, the mixed histogram is indicated) The brightness distribution of the filtered histogram is more and more similar, and the value of the mixing parameter is gradually increased. Finally, when the ratio is greater than a critical value, the mixing parameter ah is set to a constant (by 4A). For example, the ~4C graph has a constant equal to 2). However, in the application implementation, the three relationship curves have different degrees of difficulty: the stair type relationship curve (Fig. 4B) is the easiest to implement 'relatively' non- The difficulty of implementing the relationship curve (Fig. 4C) 10 200908727 degrees is the southernmost. It is assumed that the mixing parameter (1)·„ (=1 _5) is generated according to a proportional value (=maxZ^/wep=128/64 = 2, such as Figure 3 shows one of the three mixed parameters and the ratios in the 4th to 4th graphs. According to the above parameters, the weight calculator 13〇 can dynamically generate a blending weight SW. In this embodiment The weight calculator 130 uses the following mathematical equations to determine the blending weight: maxbin x in; + step x (1-^) = dgain x step...........(A) The value § ten nose device 13 0 only needs to substitute 28, and do d«=1.5 into equation (A), and can obtain the mixed weight value chat = 〇 · 5. Then, the mixing unit 140 is based on the calculated blending weight value (10) ( =〇5), the hybrid wave histogram 3 1 and the user-defined histogram 3 2, step by step to create or generate a mixed histogram 33 in a manner of bin-bin mixing. The final count value of each box type of the mixed histogram 33 can be obtained, for example, the final count of the second box type area of the mixed histogram 33 The value is equal to 96 (=128 X 0_5 + 64 X 〇·5), and the final count value of the first box type is equal to 64 (=64 X 0_5 + 64 X 0.5), etc. In equation (A), The function of _ζ>2Χ^φ) is used to limit the count value of each box type to avoid the situation of brightness distribution adjustment and over-enhancement. Finally, the 'brightness mapper 17 0 receives the mixed histogram and executes The integral operation is to generate a conversion function, which is then mapped according to the conversion function. It is to be noted that the parameter w of the present invention is not the only count value that is uniquely represented in any box type of the filter histogram and the user-defined histogram. The similarity ^ can be any value selected from a preset range, which is the maximum count value to the Nth count value in all the 200908727 type zones of the filter histogram. Where N is equal to the total number of box-type regions of the filtered histogram divided by 3 and N is a positive integer. For example, if the total number of box-type regions of the filtered histogram is equal to 15, it may be selected from any box type region of the filtered histogram, and has a maximum count value of 5 (N=15/3=5) Any one of the five count values of the large count value. In addition, the similarity / / can also be a mean value within a predetermined range 'the preset range is the third largest count value to the Nth largest count value in any of the box type regions of the filtered histogram, wherein , N is equal to the total number of box-type regions of the filtered histogram divided by 3, Μ, N are positive integers and N > M. For example: When M=2 and the total number of box-type areas of the filtered histogram is equal to 15, then the second () maximum count value of all the box-type areas of the filtered histogram seems to be 5th ( N==l5/3=5) One of the five count values of the large count value. Similarly, it may be any value within a preset range, which is a maximum count value to a maximum P count value in any box type area of the user-defined histogram, where p is equal to The total number of box-type areas of the user-defined histogram is divided by 3 and p is a positive integer. In addition, the "printing may also be a mean value within a preset range, which is the Qth large count value to the Pth large count value in all the box type areas of the user-defined histogram, wherein, P The total number of box-type areas equal to the user-defined histogram is divided by 3' Q, P is a positive integer and p > Q. Figure 5 is a schematic diagram of the architecture of the second embodiment of the present invention. The invention dynamic contrast enhancement device 5〇〇 includes a histogram calculator 110, a low-pass filter 12〇, a mixing unit 14〇, a histogram limiter 5′′, a temp〇ral filter 520, A brightness mapper 150, a normalizer 170, and a scene detector 530. Phase 12 200908727 Compared to the first embodiment, the dynamic contrast enhancement device 5 of the second embodiment lacks a weight calculator 130, but the histogram limiter 51 〇, the time domain chopper 520, and the scene detector 53 多 are added. As for the components having the same reference numerals in the second embodiment and the first embodiment, the same functions are also provided. 'Not to be repeated here. In the second embodiment, the mixing list The meta 14 利用 system uses a fixed '(Static) blending weight to blend the filtered histogram with the user-defined • histogram to gradually create a mixture (in a histogram) by mixing one box area at a time. In other words, Binb(1)=binf(1)x(10)+ Μηυ(1)χ(1-predicate), where binb(1), binf(i) and binu(1) are the count values of the mixed box histogram, the filtered histogram and the second box type in the user-defined histogram, respectively. In order to avoid the histogram brightness distribution adjustment is too intense, a histogram limiter 510 is set after the mixing unit 14〇. The histogram is limited to 5 1 0. The limit value of each box type in the mixed histogram must fall on a specific Range: (G1 X price) and (G2 X wep), where "prints represent the largest count in all box types of user-defined histograms; values, G1, G2 are positive integers and G2 > G1. That is to say, if the raw count value of any box type of the mixed straight square graph exceeds (G2 χ “卬), the final count value of the box type area will be the histogram limiter 5 1 Decrease to (G2 χ WM); conversely, if mixed If the raw count value of any box type of the histogram is lower than (G1 χ , the final count value of the box type area will be increased by the histogram limiter 510 (G1 χ 尸). Please continue to refer to Figure 5 'Scenario The detector 5 3 〇 calculates the histogram distribution gap between the two pages before and after the calculation, and compares the gap with a threshold τ. When the difference is greater than the threshold ,, the scene detector 53 determines that a field has been generated 13 200908727 Scene change (that is, in a video sequence)
中偵測到一場景轉換),並產生一致能(e n a b 1 e d)轉場訊 號;否則’場景偵測器530即產生一禁能(disabled)轉場 訊號。若轉場訊號的狀態是致能時,表示前後二頁框間的 場景有足夠大的差距,時域濾波器52〇直接輸出目前頁框 的混合直方圖;反之,若轉場訊號的狀態是禁能時,表示 該視訊序列目前在時間與空間座標上是平順且流暢的,時 域濾波器520混合目前頁框的混合直方圖與前一頁框的 混合直方圖,以一次混合一個箱型區的方式,來逐步建立一 個時域直方圖,例如,目前頁框與前一頁框係以權值分別 等於0. 25與〇· 75的方式被混合。之後,時域濾波器52〇 將該時域直方圖傳送至亮度映射器丨5〇以產生轉換函數 與執行映射處理。 第6圖為本發明第三實施例的架構示意圖◦泉考第6 圖二本。發明動態對比強化裝置6〇〇包含—局部亮度直方圖 =算器610、一低通渡波器120、一權值計算器13〇、一 σ單元14〇日守域濾波器520、—亮度映射器150、 —場:偵測器53〇、-色度映射器620以及一正規器17〇。 :三實施例和第一與第二實施例中所具有相同標號之元 件’亦具有相同之功能,在此不予贅述。 般而s,亮度直方圖中亮度最暗的箱型區,通常具 :較低::號雜訊比(signaM〇_n〇ise,)之區 I,:以二個區域中的低亮度訊號與雜訊容易混雜在- η絕If辛。根據上述之原因,最暗箱型區之對比強化 就沒有絕對的必要性 不知a之局部亮度直方圖計 14 200908727 算器6 1 0只針對一預巧·銘刑pg·於m / 孭D又相型區軏圍(例如y足第i個箱型 區到第7個箱型區,假設亮度位階總數為2⑷,產生一 局部亮度直方圖,而不去計算第〇個箱型區(即,,勿、視” 區域)的計數值。接著’低通遽波器12〇對該局部亮度直 方圖執行低通濾波處理,以產生一濾波直方圖。在執行正 規化處理之前,正規器17Q對所有輸人的頁框的濾波直方 圖,都設定(或限制)帛。個箱型區的計數值等於一預設 常數F,。之後’正規器17Q對於像素總數目為心⑷之A scene transition is detected and a consistent energy (e n a b 1 e d) transition signal is generated; otherwise, the scene detector 530 generates a disabled transition signal. If the state of the transition signal is enabled, it means that the scene between the two pages is sufficiently wide, and the time domain filter 52 directly outputs the mixed histogram of the current page frame; otherwise, if the state of the transition signal is When disabled, it indicates that the video sequence is currently smooth and smooth on the time and space coordinates. The time domain filter 520 mixes the mixed histogram of the current page frame with the mixed histogram of the previous page frame to mix one box at a time. The way of the area, to gradually establish a time domain histogram, for example, the current page frame and the previous page frame are mixed with a weight equal to 0.25 and 〇 75 respectively. Thereafter, the time domain filter 52 传送 transmits the time domain histogram to the luminance mapper 丨5〇 to generate a conversion function and perform mapping processing. Figure 6 is a schematic diagram of the architecture of the third embodiment of the present invention. The invention dynamic contrast enhancement device 6〇〇 includes a local brightness histogram=a calculator 610, a low-pass waver 120, a weight calculator 13〇, a sigma unit 14 守 day-domain filter 520, a brightness mapper 150, - field: detector 53 〇, - chromaticity mapper 620 and a regularizer 17 〇. The third embodiment has the same functions as those of the first and second embodiments, and the same functions are not described herein. Generally, the darkest box area in the luminance histogram usually has a lower::noise ratio (signaM〇_n〇ise) area I,: a low-brightness signal in two areas It is easy to mix with noise - η 绝 If. According to the above reasons, there is no absolute necessity for the contrast enhancement of the darkest box type. I don't know the local brightness histogram of a. 14 200908727 Calculator 6 1 0 only for a pre-exception · Ming pg · in m / 孭D The area of the type zone (for example, the i-th box type to the seventh box type area of the y foot, assuming that the total number of brightness levels is 2 (4), generating a partial brightness histogram without calculating the second box type area (ie, Do not look at the count value of the region. Then the low pass chopper 12 执行 performs low pass filtering on the local luminance histogram to generate a filtered histogram. Before performing the normalization process, the normalizer 17Q pairs all The filtered histogram of the input frame is set (or limited). The count value of the box type is equal to a preset constant F. After that, the regularizer 17Q is the heart for the total number of pixels (4).
濾波直方圖,將預設箱型區範圍(第】個 箱型…的計數值正規化至…。上述的」= 式不但可以避免晝面之黑邊(black ba〇干擾,而且也能 避免產生亮度對比過度強化之效果,此外,固定最暗箱型 區的計數值還能改善使用者對亮度對比的感知,並避免曝 光過度(washout)的問題。理論上,指定給,,忽視”區域 j第〇個箱型區)的常數Fl值越大,則可供預設箱型區 範圍(第1個箱型區到第7個箱型區)分配的剩餘亮度位 P白數目就越少。凊注思,”忽視”區域並不限於第〇個箱型 區,同時,常數值大小與預設箱型區範圍大小皆可根 據需求來修改。 、根據第三實施例,亮度資訊還能以不同於先前技術的 方式來強化色度對比。根據轉換函數與一輸入亮度位階 X,亮度映射器150更產生一色度增益(gain)。確切而言, 亮度映射器150在產生轉換函數之後,接收—輸入像素 (或—輸入影像資料)之亮度成份X並產生一相對應的輪 15 200908727 出亮度位階Κ的方式,係以每一個亮度位階為單位(^ ieve〗 by level basis)。接著,亮度映射器15〇對該輸入像 素的亮度增益進行修改,以產生該輪入像素之亮度 成份的色度增益’色度增益之產生係以每—個箱型區為單 位(on a bin by bin basis)。據此,色度映射器62〇即可 將該色度增益運用到(apply)該輪入像素之一輸入亮度成 份。 以下第7A圖與第7B圖的例子係用來說明亮度映 射器15〇與色度映射器620之運作。第7a圖顯示時二直 方圖與其相對應轉換函數的一個例子。第7B圖係根據第 7A圖的轉換函數,顯示色度增益與箱型區索引之間的一 個關係圖。 第7A圖的時域直方圖中,假設L(亮度位階總數) =256>以及bsize(總箱型區數)=8時,則每一箱型區寬度 =32焭度位階。在接收完時域直方圖之後,亮度映射器 將bin⑴積分就能得到轉換函數= ⑺,其 中cwrve⑼-〇 ’ bln⑴是時域直方圖第丨個箱型區的計 婁值 為正JE'數並涵盍1到(bsize-Ι)的範圍。實際應用 時’轉換函數㈣v咐ϋ多段式線性函數,在第'A 圖/以實線表示’每—箱型區均具有—相對應的直線線 &而母—直線線段之斜率函數m⑺=(bin⑴/, 其中’丨為正整數並涵蓋1到(bsize-Ι)的範圍。 根據第二貫施例,輪出亮度位階&是由亮度映射器 16 200908727 150根據相對應的輸入亮度位階Γ,及二個函數 m(/)而產生。輸出亮度位階ι以數學方程式表示如下:I = (wrv#办;+ ο//ίβί χηφ办)),其中,z.办=fi〇〇ru/ww^), OetmodUAw^/O, 是小於或等於之最接 近的整數,而是的餘數。由於人的肉眼 對色度比較不敏感,因此色度增益的產生係以箱型區為單 位,進而避免色彩對比的過度強化。如此,亮度映射器 15〇根據轉換函數cwrve⑺及一輸入像素(或一輸入影像 料)之壳度位階K,對該輸入像素之色度成份,產生一 色度增益。色度增益以數學方程式表示如下: cg(〇=⑶rvw+u/(w/c^x(出)),其中,w(bsize_1)=1,,·為 一整數並涵蓋0到(bsize-2)的範圍。例如,當輸入亮度位 階K在第0個箱型區的範圍内時,色度增益cg(〇)= cwrve〔lj/( iWi/Mxl));而當輸入亮度位階y在第j個箱型 區的範圍内時,色度增盈c^(l)= 心Αχ2)), 以此類推。最後,色度映射器620將該色度增益運用於該 輸入像素之一輸入色度成份ς ’以得到相對應的輸出色度 成份C„ = C, X ,其中,,必是 小於或等於之最接近的整數。 综上所述’本發明為直方圖等化法之改良,依據每個 影像產生一對一的轉換函數,而不需要從一組預設轉換函 17 200908727 數t挑選, 之缺點。 並且避免產生調整過劇 α &平均亮度偏移過大 隹較佳實施例之詳 m ^ .....〜山 < 具體賞施例僅 制明本發明之技術内容’而非將切明狹義地限 範圍3Γ:在不超出本發明之精神及以下申請專利 做之種種變化實施,皆屬於本發明之範圍。 18 200908727 【圖式簡單說明】 第1圖為本發明第一實施例的架構示意圖。 第2圖是兩個轉換函數的比較圖。 第3圖係顯示混合一濾波直方圖與—使用者自訂直方 圖’以產生一混合直方圖的一個例子。 第4A〜4C圖顯示三條不同的混合參數與比例之關係 曲線。 ” 第5圖為本發明第二實施例的架構示意圖。 f 第6圖為本發明第三實施例的架構示意圖。 第7A圖顯示時域直方圖與其相對應轉換函數的一個 例子。 第圖係根據第7A圖的轉換函數,顯示色度辦兴 與箱型區索引之間的__個關係圖。 【主要元件符號說明】Filter the histogram to normalize the count value of the preset box type range (the first box type...). The above "=" can not only avoid the black side of the face (black ba〇 interference, but also avoid it. The brightness contrast is over-enhanced. In addition, the count value of the fixed darkest box area can also improve the user's perception of brightness contrast and avoid the problem of overshoot. In theory, assign to, ignore the area j The larger the constant Fl value of one box type area, the smaller the number of remaining brightness bits P white that can be allocated for the preset box type area (1st box type to 7th box type area). Note that the "ignore" area is not limited to the second box type area, and the constant value size and the preset box type area size can be modified according to requirements. According to the third embodiment, the brightness information can be different. The chromaticity contrast is enhanced in a prior art manner. The luminance mapper 150 produces a chrominance gain based on the conversion function and an input luminance level X. Specifically, the luminance mapper 150 receives the conversion function and receives - lose The brightness component X of the pixel (or - input image data) and a corresponding wheel 15 200908727 brightness level Κ is obtained in units of each brightness level (^ ieve by level basis). Then, the brightness mapper The luminance gain of the input pixel is modified to produce a chrominance gain of the luminance component of the rounded pixel. The chrominance gain is generated on a bin by bin basis. Accordingly, the chrominance mapper 62 can apply the chrominance gain to one of the input luminance pixels of the rounded pixel. The following examples of FIGS. 7A and 7B are used to illustrate the luminance mapper 15〇. Operation with chroma mapper 620. Figure 7a shows an example of a two-square histogram and its corresponding transfer function. Figure 7B shows the conversion between the chroma gain and the bin index based on the transfer function of Figure 7A. A diagram. In the time domain histogram of Fig. 7A, assuming L (the total number of luminance levels) = 256 > and bsize (the total number of bins) = 8, then each box type width = 32 degrees. After receiving the time domain histogram, The degree mapper integrates bin(1) to get the conversion function = (7), where cwrve(9)-〇' bln(1) is the time-domain histogram. The second box type has a positive JE' number and covers 1 to (bsize-Ι). The range of the actual application 'transfer function (four) v 咐ϋ multi-segment linear function, in the 'A figure / by the solid line indicates that 'every box type has - the corresponding straight line & and the mother - the slope of the straight line segment The function m(7)=(bin(1)/, where '丨 is a positive integer and covers the range of 1 to (bsize-Ι). According to the second embodiment, the rounded luminance level & is determined by the brightness mapper 16 200908727 150 according to The luminance level Γ is input and two functions m(/) are generated. The output brightness level ι is expressed as a mathematical equation as follows: I = (wrv# do; + ο// ίβί χηφ)), where z.do =fi〇〇ru/ww^), OetmodUAw^/O, is less than or Equal to the nearest integer, but the remainder. Since the human eye is less sensitive to chromaticity, the chromatic gain is generated in a box-type area, thereby avoiding excessive enhancement of color contrast. Thus, the luminance mapper 15 generates a chrominance gain for the chrominance components of the input pixel based on the conversion function cwrve (7) and the shell degree K of an input pixel (or an input image). The chrominance gain is expressed as a mathematical equation as follows: cg(〇=(3)rvw+u/(w/c^x(out)), where w(bsize_1)=1, ,· is an integer and covers 0 to (bsize-2 For example, when the input luminance level K is within the range of the 0th box type, the chrominance gain cg(〇) = cwrve[lj/(iWi/Mxl)); and when the input luminance level y is at the When the range of j box type is within, the chrominance gains c^(l)= palpitations 2)), and so on. Finally, the chroma mapper 620 applies the chroma gain to one of the input pixels to input a chroma component ς ' to obtain a corresponding output chroma component C„=C, X , where, must be less than or equal to The most recent integer. In summary, the present invention is an improvement of the histogram equalization method, which generates a one-to-one conversion function according to each image, without selecting a set of preset conversion letters 17 200908727 Disadvantages. And avoiding the adjustment of the drama α & average brightness shift is too large 隹 详 隹 隹 隹 隹 ^ ^ ^ ^ ^ ^ ^ ^ ^ 具体 具体 具体 具体 具体 具体 具体 具体 具体 具体 具体 具体 具体 具体 具体 具体 具体 具体 具体 具体It is to be understood that the scope of the present invention is not limited by the spirit of the present invention and the various modifications of the following patent application. 18 200908727 [Simple Description of the Drawings] FIG. 1 is a first embodiment of the present invention. Schematic diagram of the structure. Figure 2 is a comparison of two conversion functions. Figure 3 shows an example of a mixed-filter histogram and a user-defined histogram to generate a mixed histogram. Figures 4A to 4C show Three different Relationship between the ratio of the mixing parameter curve. "5 architecture diagram of the second graph according to the present embodiment of the present invention. f Figure 6 is a block diagram showing the architecture of the third embodiment of the present invention. Figure 7A shows an example of a time domain histogram and its corresponding conversion function. The figure shows a __ relationship diagram between the chroma run and the box type index according to the transfer function of Fig. 7A. [Main component symbol description]
6〇〇動態對比強化裝置 120低通濾波器 140混合單元 160目前頁框 100 ' 500 11 〇直方圓計算器 13 0權值計算器 15 0亮度映射器 170正規器 22、24轉換函數 32使用者自訂直方圖 5 1 〇直方圖限制器 5 3 0場景偵測器 620色度映射器 21亮度直方圖 23、31濾波直方圖 33混合直方圖 5 20時域濾波器 610局部直方圖計算器 196〇〇 dynamic contrast enhancement device 120 low pass filter 140 mixing unit 160 current page frame 100 '500 11 〇 straight square calculator 13 0 weight calculator 15 0 brightness mapper 170 regular device 22, 24 conversion function 32 user Custom Histogram 5 1 〇 Histogram Limiter 5 3 0 Scene Detector 620 Chroma Mapper 21 Brightness Histogram 23, 31 Filter Histogram 33 Mixed Histogram 5 20 Time Domain Filter 610 Local Histogram Calculator 19
Claims (1)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW96129495A TW200908727A (en) | 2007-08-10 | 2007-08-10 | Method and apparatus for dynamic contrast enhancement |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| TW96129495A TW200908727A (en) | 2007-08-10 | 2007-08-10 | Method and apparatus for dynamic contrast enhancement |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| TW200908727A true TW200908727A (en) | 2009-02-16 |
Family
ID=44723747
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| TW96129495A TW200908727A (en) | 2007-08-10 | 2007-08-10 | Method and apparatus for dynamic contrast enhancement |
Country Status (1)
| Country | Link |
|---|---|
| TW (1) | TW200908727A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI497320B (en) * | 2011-11-14 | 2015-08-21 | Novatek Microelectronics Corp | Automatic tone mapping method and image processing device |
| TWI668668B (en) * | 2018-05-28 | 2019-08-11 | 多方科技股份有限公司 | Method and computer system of image enhancement |
-
2007
- 2007-08-10 TW TW96129495A patent/TW200908727A/en unknown
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI497320B (en) * | 2011-11-14 | 2015-08-21 | Novatek Microelectronics Corp | Automatic tone mapping method and image processing device |
| TWI668668B (en) * | 2018-05-28 | 2019-08-11 | 多方科技股份有限公司 | Method and computer system of image enhancement |
| US10628929B2 (en) | 2018-05-28 | 2020-04-21 | Augentix Inc. | Method and computer system of image enhancement |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109937444B (en) | display device | |
| CN105282452B (en) | Image processing apparatus and image processing method | |
| CN105009580B (en) | The display management of high dynamic range video | |
| CN109983530A (en) | Ambient light adaptive display management | |
| CN107211076B (en) | The method, apparatus and storage medium of display management for high dynamic range video | |
| US20180124368A1 (en) | Color gamut mapper for dynamic range conversion and methods for use therewith | |
| EP3280138A2 (en) | Image processing apparatus, image projection apparatus, and image processing method | |
| US20150256860A1 (en) | Graphics Blending for High Dynamic Range Video | |
| JP5439588B2 (en) | Apparatus and method for processing image data for display on a display panel | |
| CN107888943B (en) | Image processing | |
| CN102752604B (en) | Image display method and intelligent device | |
| US20170124983A1 (en) | Adaptive Display Management Using 3D Look-Up Table Interpolation | |
| TW201026080A (en) | Method and system for scene adaptive dynamic 3-D color management | |
| CN112750086A (en) | Image processing method and device, electronic equipment and storage medium | |
| CN105323502A (en) | Image processing apparatus and image processing method | |
| JP2018026794A (en) | Image processing system, image projection device, and image processing method | |
| CN108141576A (en) | Display device and control method thereof | |
| TWI361003B (en) | Digital image tone remapping method and apparatus | |
| JP2014220815A (en) | Image color adjusting method and electronic device using the same | |
| Song et al. | Luminance enhancement and detail preservation of images and videos adapted to ambient illumination | |
| TW200908727A (en) | Method and apparatus for dynamic contrast enhancement | |
| JP4443406B2 (en) | Non-linear image processing | |
| JPWO2018003665A1 (en) | Image processing apparatus, image processing method, control program, and recording medium | |
| CN111247554A (en) | Image processing apparatus, image processing method, and image processing program | |
| JP2002158894A (en) | Dynamic gamma correction device |