TW391130B - Method and apparatus for contrast enhancement of color images - Google Patents
Method and apparatus for contrast enhancement of color images Download PDFInfo
- Publication number
- TW391130B TW391130B TW85110136A TW85110136A TW391130B TW 391130 B TW391130 B TW 391130B TW 85110136 A TW85110136 A TW 85110136A TW 85110136 A TW85110136 A TW 85110136A TW 391130 B TW391130 B TW 391130B
- Authority
- TW
- Taiwan
- Prior art keywords
- value
- color
- contrast
- sampling
- item
- Prior art date
Links
Landscapes
- Image Processing (AREA)
Abstract
Description
經濟部中央樣準局員工消費合作社印裝 A7 B7__ 五、發明説明(/) 1 本發明偽有閨於一種彩色影像對比加強之方法及其《 路裝置,待別像有鬭於一種解壓缩後之彩色彩像,利用其 在HSI色彩空間之對應I值,産生原彩色彩像對比加強之 方法及其霣路装置。 5 按,為了減少影像傳输所需頻宽或皤存所霈的記慵韻 ,各種彩像壓缩技術被廣泛地使用。在壓缩及解壓缠的遇 程中,往往會播牲部份高頻訊號以逋成歷缩效率及即時解 壓縮,而使得解壓缩後的彩像頭得棋糊,色彩暗淡,逭時 應用對比加強的技巧躭能重造色彩鮮艷生勡的影像。 10 對黑白影像來說,對比加強的技巧可利用拉大彩像亮 度(luninance)的差異範園來增強對比;然而,對彩色影 像除了亮度外邇要考«顔色的飽和度(saturation)及色度 (hue)。 此外,同一種顔色在不同的色彩空間(color space) <- --- 15 有對應的空間座標值,因此要遘擇在何種色彩空間上作對 比加強也是門學問。例如,在Y Cr Cb之色彩空間上加強 對比時,發現當對比加強的乘數(contrast enhance factor) 加大時,會使得彩像的亮度提离,色彩匏和度減低。 有鑑於此,本發明人基於傳統彩色彩像對比加強之未 20 臻理想,在兼顧整鼸設計登實用下稹棰研究改良,遂有本 發明彩色彩像對比加強之方法及其«路裝置産生。 邸,本發明之第一目的,在於提供一種利用原彩色彩 像在HSI色彩空間之對應I值逢成彩色影像對比加強之方 法及其電路裝置。 _第04頁_. 本紙張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) (請先閲讀背面之注意事項再填寫本頁) 裝· 訂 經濟部中央樣準局員工消費合作社印裝 A7 B7 五、發明説明(2 ) 1 本發明之第二目的,在於提供一種彩色彩像對比加強 之方法及其轚路裝置,對原彩色彩像之像素進行取樣,以 減低硬匾之成本。 本發明之第三目的,在於提供一種彩色彩像對比加強 5 之方法及其霣路裝置,除可對靜態的單一彩像畫面進行對 比加強外,更可對動態的畫面進行對比加強者。 緣是,為建上述目的,本發明之第一特微你在於包括 一第一計算單元,以計算解壓编後彩色膨像之毎餹像素資 料在HSI色彩空間中對應之I值;一第二計算單元,以計 10 算所有I值,输出一新I值;及一對比加強運算單元,以 利用新I值搿比加強解壓缩後彩色彩像的毎鵃像素資料, 使影像在亮度及色彩飽和度得以改菩。 本發明之第二特微你在於更包括一取樣運算單元,以 、 取樣上述解壓缩後彩色影像之部份像素資料i上述第一計 15 箄單元,該取樣蓮算單元包括有一第三計算單元、一比較 器、一計數器、及一取樣黏數比較器,其中該第三計算單 元包括有一累加器,用以毎讀取一餹像素即累加一預設之 取樣參數(大於零之整數),而輪出一累加值,及一減法 器;該比較器,其參考值則為一預設之計數最大值,且當 20 比較某點像素之累加值大於該參考值時,邸输出一取樣倍 號至上述第一、第三計算單元,該第一計算單元邸計算出 該黏像素之對應I值,且第三計算箪元之減法器邸將該累 加值減去參考值,並自差值繼鑛累加取櫬參數;該計數器 ,每接受一次比較器输出之取樣倍號,即累加—1",且 __第05頁_ 本纸張尺度適用中國國家標準(CNS ) A4規格(210X297公釐) ϊ 1 ^^1 n In— m 1 ί I— 1 n 1^1 (請先閲讀背面之注意事項再填寫本頁) 經濟部中夬標準局員工消費合作社印製 A7 B7_ 五、發明説明() 1 输出一累加值至取樣點數比較器,該取樣點數比較器參考 值則為一預設之取樣點數目,當上述計數器輸出之累加值 等於取樣黏數目時即輪出一結束取樣信號至上述第一計算 單元,以停止取樣黏之蓮算,而將所有取樣點之對應I值 5 输入第二計算單元,以在不滅亮度及色彩飽和度之效果下 ,減少硬體所需成本者。 本發明之第三特欲像在於上述第二計算單元内更包括 一暫存器及一權數乘加單元,而以該暫存器用來暫存前幾 張彩像的新I值,然後再經權數乘加單元將逭些新I值乘 10 上適當的權數後相加後做為下一張彩像對比加強的新I值 ,達到動戆影像對比加強之目的效果者》 茲配合鼷示,将本發明彩色影像鲟比加強之方法及其 霣路裝置之較佳實施例,詳細説明如下,以期能對本發明 之優點,目的及待揪,能獲得進一步之充分瞭解。 15 第一圔為本發明彩色影像對比加強之方法及霣路装置 之«路方塊_。 第二圓為第一圖中第一計算單元之第一實施例《路方 r-*' 塊_〇 第三B為第一中第一計算單元之第三實施例霣路方 20 塊疆。 第四為第一_中第二計算單元計算中值之霣路方塊A7 B7__ printed by the Consumer Cooperatives of the Central Bureau of Procurement of the Ministry of Economic Affairs V. Description of the Invention (/) 1 The invention is based on a method for enhancing the contrast of a color image and its "road device. The color image of the color image uses its corresponding I value in the HSI color space to generate a method for enhancing the contrast of the original color image and its road device. 5 Press. In order to reduce the bandwidth or memory required for image transmission, various color image compression techniques are widely used. In the process of compression and decompression, some high-frequency signals are often broadcasted to achieve historical compression efficiency and instant decompression, so that the decompressed color picture head is blurred, the color is dim, and contrast is often applied. Enhanced techniques will not recreate brightly colored images. 10 For black-and-white images, the technique of contrast enhancement can make use of widening the difference in luminosity of the color image to enhance the contrast; however, for color images, in addition to the brightness, «saturation and color Degree (hue). In addition, the same color has corresponding spatial coordinate values in different color spaces <---- 15, so it is also a matter of choosing which color space to compare and strengthen. For example, when the contrast is enhanced in the color space of Y Cr Cb, it is found that when the contrast enhancement factor is increased, the brightness of the color image is lifted off, and the color intensity is reduced. In view of this, the present inventor based on the traditional ideal of color image contrast enhancement, and researched and improved the design while taking into account the overall design. Therefore, the method for color image contrast enhancement of the present invention and its «road device production» . The first object of the present invention is to provide a method and circuit device for enhancing the contrast of a color image using the corresponding I value of the original color image in the HSI color space. _Page04_. This paper size applies to China National Standard (CNS) A4 (210X297mm) (Please read the precautions on the back before filling this page) A7 B7 V. Description of the invention (2) 1 The second object of the present invention is to provide a method for enhancing the contrast of a color image and a road device thereof, sampling the pixels of the original color image to reduce the cost of a hard plaque. A third object of the present invention is to provide a method for color image contrast enhancement 5 and a roadway device thereof. In addition to contrast enhancement of a static single color image frame, it is also capable of contrast enhancement of a dynamic image. The reason is that in order to achieve the above purpose, the first feature of the present invention is that it includes a first calculation unit to calculate the I value corresponding to the H pixel color space of the pixel data of the color expansion image after decompression; a second A calculation unit that calculates all I values by 10 and outputs a new I value; and a contrast enhancement operation unit to enhance the pixel data of the decompressed color image by using the new I value / ratio, so that the image is in brightness and color Saturation can be changed. The second special feature of the present invention is that it further includes a sampling operation unit for sampling the pixel data of the decompressed color image. The above-mentioned first counting unit is 15 mm, and the sampling calculation unit includes a third calculation unit. , A comparator, a counter, and a sampling sticky number comparator, wherein the third calculation unit includes an accumulator for accumulating a preset sampling parameter (an integer greater than zero) by reading a pixel, An accumulative value and a subtractor are rotated out; the reference value of the comparator is a preset maximum count value, and when the accumulated value of the pixels at a certain point of comparison 20 is greater than the reference value, the sampling output is doubled. To the first and third calculation units described above, the first calculation unit calculates the corresponding I value of the sticky pixel, and the subtractor of the third calculation unit subtracts the reference value from the accumulated value and subtracts the difference value. Following the mining accumulation, the parameter is incremented; this counter is incremented every time it receives the sampling multiple of the comparator output, and __page05_ This paper size applies to the Chinese National Standard (CNS) A4 specification (210X297) %) 1 ^^ 1 n In— m 1 ί I— 1 n 1 ^ 1 (Please read the notes on the back before filling in this page) Printed by A7 B7_, Consumer Cooperatives of the China Standards Bureau of the Ministry of Economic Affairs V. Description of the invention () 1 Output an accumulated value to the sampling point number comparator. The reference value of the sampling point comparator is a preset number of sampling points. When the accumulated value of the counter output is equal to the number of sampling sticks, an end sampling signal is output to the above. The first calculation unit is to stop sampling the sticky lotus, and input the corresponding I value 5 of all sampling points to the second calculation unit, in order to reduce the cost of hardware under the effect of constant brightness and color saturation. The third special image of the present invention is that the second calculation unit further includes a register and a weight multiply-add unit, and the register is used to temporarily store the new I values of the previous color images, and then the The weight multiplication and addition unit will multiply these new I values by 10 with appropriate weights and add them as the new I value for the next color image enhancement to achieve the effect of enhancing the contrast of the image. The preferred embodiment of the method for enhancing the color ratio of the color image of the present invention and the roadway device thereof will be described in detail in the hope that the advantages, objectives, and requirements of the present invention can be fully understood. 15 The first method is the method of enhancing the contrast of color images and the device of the present invention. The second circle is the first embodiment of the first calculation unit in the first figure "Road side r- * 'block_〇 The third B is the third embodiment of the first calculation unit in the first road unit. The fourth is the Kushiro block of the first _ second calculation unit that calculates the median
第五疆為一 4X4矩陣,其内具有毎一像素所對應之I 值〇 ___第06頁_ 本紙張尺度適用中國國家標準(CNS ) A4規格(2丨〇父297公釐) I- n^l — nn —^ϋ — n ml nn ^ (請先閲讀背面之注意事項再填寫本頁) 五、發明説明(斗) 15 A7 B7 第六·為第五隱中毎籲不同I值之像素數目的分佈曲 線鼷。 第七·為第五麵中之像素在每餹I值内的累計分佈曲 鎳疆。 第八·為第一中第二計算單元計算平均值之《路方 塊圓。 _九鼷為一適用於本發明之取樣霣路方壤·。 第+為本發明彩色影像對比加強之方法及霣路装置 之一較佳實施例霣路方塊·。 第十一_為第一、+圔中對比加強邇算單元之較佳實 施例《路方塊·。 第十二為第十匾變化實施於動態像$霣路方塊匯. (請先閲讀背面之注意事項再填寫本頁) 裝- -6 經濟部中央樣準局貝工消費合作社印裝 20 鼷號對照: 1第一計算單元 10 除法器 12 計數器 2第二計算單元 20中值計算器 22 平均值計算器 24 第二暫存器 3對比加強麵》單元 30 差值計算器 32 加法器 11 累加器 13 多工器 21暫存器 23 暫存器 25 權數乘加單元 31對比加強增益蓮算器 \ 第07頁 本纸張尺度適用中國國家梯準(CNS ) A4規格(210X297公釐) 經濟部中央樣準局貝工消費合作社印裝 A7 B7 五、發明説明(s) 1 4取樣邐算單元 40 第三計算單元 41比較器 42 計數器 43 取樣點數比較器 讅參閲第一所示,為本發明彩色彩像獬比加強之方 5 法及《路裝置之《路方塊匾,包括有一第一計算單元1, 以計算解壓编後彩色彩像之毎餹像素資料在HSI色彩空間 中對應之I值;一第二計算單元2,以計算上述所有I值 ,以輪出一新I值;及一對比加強蓮算單元3,接受上述 新I值,以將解壓縮後彩色影像的毎僮像素資料作對比加 10 強。 其中,若解S缠後之彩色彩像供屬RgB色彩空間,則 HSI色彩空間與RGB色彩空間的轉換B像式即如下所示: H=C0S-Ml/2 [(R-G) + (R-B)]/ [ (R-β)2 + (R-B) (G-B) ] } ............................................(1) 15 S=l-3/(R + G + B) [ain(R,Gf B)].....................(2) I=(R + G + B)/3. . ................................(3) 其中:H表示色度;S表示飽和度;I表示亮度。 上述第一計算單元1則是用來計算上式(3),而在本 發明之^實施例中,謫參開第二所示,該Ϊ一計算單 20 元1包括有一除法器1〇,以將同一像素資料之R、G、B值 除以3而輪出R/3, G/3, B/3值;及一累加器11,將上迷 R/3,G/3, B/3值相加而输出(R+G+B)/3 ,邸I值。 第一計算單元1之第二實施例供將第二之除法器改 用一計算器,将同一像素賫料之R、G、B值分別以(1/4 + 第08頁 本紙張尺度適用中國國家標準(CNS ) Α4规格(210X297公釐) - 1 i— I ITU In H^I ! i I I I (資先閲讀背面之注意事項再填寫本頁)The Fifth Xinjiang is a 4X4 matrix, which has an I value corresponding to one pixel. ______Page 06_ This paper size applies the Chinese National Standard (CNS) A4 specification (2 丨 〇 parent 297 mm) I- n ^ l — nn — ^ ϋ — n ml nn ^ (Please read the notes on the back before filling out this page) V. Description of the invention (bucket) 15 A7 B7 Sixth: Call for different I values in the fifth cover The distribution curve of the number of pixels 鼷. Seventh is the cumulative distribution curve of the pixels in the fifth plane within each 餹 I value. Eighth: The road square block circle which calculates the average value for the first and second calculation units. _ Jiuyi is a sampling Kushiro soil suitable for the present invention. The first + is a preferred embodiment of a method and a device for color contrast enhancement of a color image of the present invention. The eleventh _ is a better embodiment of the first and + contrast-enhanced computing units "Road Blocks." The twelfth is the tenth plaque change implemented on the dynamic image $ 霣 路 方 汇. (Please read the precautions on the back before filling out this page) Pack--6 Printed by the Central Laboratories Bureau of the Ministry of Economic Affairs Contrast: 1 first calculation unit 10 divider 12 counter 2 second calculation unit 20 median calculator 22 average calculator 24 second register 3 contrast enhancement surface "unit 30 difference calculator 32 adder 11 accumulator 13 Multiplexer 21 Register 23 Register 25 Multiply and add unit 31 Contrast gain gain calculator \ Page 07 This paper size is applicable to China National Standard (CNS) A4 (210X297 mm) Central Ministry of Economic Affairs Printed by the Bureau of Specimen Cooperative Consumer Cooperatives A7 B7 V. Description of the invention (s) 1 4 Sampling calculation unit 40 Third calculation unit 41 Comparator 42 Counter 43 Sampling point comparator Invented the method 5 for enhancing the color ratio of color images and the road block plaque, which includes a first calculation unit 1 to calculate the corresponding I of the pixel data of the decompressed color image in the HSI color space. Value; a second calculation unit 2, In order to calculate all the above-mentioned I-values, a new I-value is calculated; and a contrast-enhancing lotus calculation unit 3, which accepts the above-mentioned new I-values, in order to compare and strengthen the child pixel data of the decompressed color image with a strong 10. Among them, if the color image after the unwrapping is for the RgB color space, the conversion of the HSI color space to the RGB color space is as follows: H = C0S-Ml / 2 [(RG) + (RB) ] / [(R-β) 2 + (RB) (GB)]} ... ............ (1) 15 S = l-3 / (R + G + B) [ain (R, Gf B)] ............ ......... (2) I = (R + G + B) / 3 ... ....... (3) where: H represents chroma; S represents saturation; I represents brightness. The first calculation unit 1 is used to calculate the above formula (3), and in the embodiment of the present invention, as shown in the second, the first calculation unit 20 yuan 1 includes a divider 10, The R / 3, G / 3, and B / 3 values are rounded out by dividing the R, G, and B values of the same pixel data by 3; and an accumulator 11 will add R / 3, G / 3, B / Add the 3 values and output (R + G + B) / 3, the value of I. The second embodiment of the first calculation unit 1 is used to change the second divider to a calculator, and the R, G, and B values of the same pixel data are divided by (1/4 + page 08) National Standard (CNS) Α4 Specification (210X297 mm)-1 i— I ITU In H ^ I! I III (Please read the notes on the back before filling this page)
*1T 丨外 經濟部中央標準局貝工消費合作社印装 A 7 B7 五、發明説明(6 ) 1 1/16 + 1/64----)之運算遍近而輪出R/3, G/3, B/3值至上 述累加器11輪出(R+G+B)/3。 而第一計算單元1之第三實施例,讅參閲第三鬭所示 ,包括一毎次數到2又會再從0两始數之計數器12,以输 5 出〇、1、2三籲控制倍號;一多工器13,分別接受三值連 鑛儋素資料(Rl、Gl、B1)(R2、G2、B2)(R3、G3、B3)及上 述控制倍«,在本較佳*施例中,當該多工器13接,受第一 控制倍》| 〇#時輪出第一鑌像素賫料(R1、Gl、B1)的R1 值,接受第二控制信» " 1"時輪出第二《像索(R2、G2 10 、B2)的G2值,及接受第三控制倍虢* 2 #時輪出第三梅 像素資料(R3、G3、B3)的B3值;及上述累加器11,將R1、 G2、B3三值相加,因為影像的待性使鄰近像素的R、G、B 值幾乎相同或僅有缓慢的增《,故R1 + G2 + B3之累加值, [1/3(R1 + G1 + B1) + 1/3(R2 + G2 + B2) + 1/3(R3 + G3 + B3)]之累加 15 值。 再者,上述第二計算器2你用來取得上述所有像素I 值之中值或平均值;其中:中倌取法如下: r — I > 1 . 一串數列的中值(median)定義為在逭串數列中半數的 數值小於逭值中值,而另外一半的數值則大於中值。根據 20 Philip E.Mattison的描述,影像的對比加強可依下列式 子達成:* 1T 丨 Printed by the Central Standards Bureau of the Ministry of Foreign Affairs, Shellfish Consumer Cooperatives A 7 B7 V. Description of the invention (6) 1 1/16 + 1/64 ----) The calculations are rounded up and R / 3, G are rotated out / 3, B / 3 value to the totalizer 11 round out (R + G + B) / 3. The third embodiment of the first calculation unit 1 is shown in FIG. 3, and includes a counter 12 that counts from 2 to 2, and then counts from 0 to 2, and outputs 5 outputs of 0, 1, 2, and 3. Control multiplier number; a multiplexer 13 that accepts three-value continuous ore element data (Rl, Gl, B1) (R2, G2, B2) (R3, G3, B3) and the aforementioned control multiplier «, better in this book * In the example, when the multiplexer 13 is connected, it is subject to the first control times. | 〇 # The R1 value of the first pixel data (R1, Gl, B1) is rounded off, and the second control letter is accepted »" 1 " Time to get the second "G2 value of the image cable (R2, G2 10, B2), and accept the third control time 虢 * 2 #Time to show the third Mei pixel data (R3, G3, B3) B3 value ; And the accumulator 11 above, the three values of R1, G2, and B3 are added, because the R, G, and B values of adjacent pixels are almost the same or only increase slowly because of the waiting property of the image, so R1 + G2 + B3 Accumulated value, the value of [1/3 (R1 + G1 + B1) + 1/3 (R2 + G2 + B2) + 1/3 (R3 + G3 + B3)] is added to 15 values. Furthermore, the second calculator 2 you used to obtain the median or average value of the I values of all the pixels; where: the median selection method is as follows: r — I > 1. The median of a series of numbers is defined as Half of the values in the 逭 series are less than the median of 逭, and the other half are greater than the median. According to 20 Philip E. Mattison, the contrast enhancement of the image can be achieved by the following formula:
Pixel_Value = (Pixel_Value - Median_Value) XPixel_Value = (Pixel_Value-Median_Value) X
Contrast.Enhanceaent.6ain + Median.Value 如上所述,當使用在彩色彩像時,本發明将之修正為 _第09頁_ 本紙張尺度適用中國國家標準(CNS ) A4洗格(210X 297公釐) I- — - - «1— el^i -=r- — I I «II is -- — —-II ('請先閲讀背面之注意事項再填寫本頁) 五、發明説明(7 ) 15 A7 B7 經濟部中央橾準局員工消费合作社印装 20Contrast.Enhanceaent.6ain + Median.Value As mentioned above, when used in color images, the present invention will modify it to _ page 09 _ This paper size applies the Chinese National Standard (CNS) A4 Washing (210X 297 mm) ) I- —--«1— el ^ i-= r- — II« II is-— —-II ('Please read the notes on the back before filling out this page) 5. Description of the invention (7) 15 A7 B7 Printed by the Consumers' Cooperatives of the Central Government Bureau of the Ministry of Economic Affairs 20
Median.I Median of I Enhanced_R=(R-Median_I)xContrast_Enhance«ent_Gain + Median_I...........................(4) Enhanced_G=(G-Median_I)x Contrast_Enhanceient_6ain + Median_I...........................(5) Enbanced_B=(B-Nedian_I)xContrast_Enhanceaent_6ain + Median_I...........................(6) 而要計算一張彩像的中值,請參閲第四所示,上述 第二計算單元2包括一中值計算器20及一轚存器21,而以 該中值計算器20统計像素(Pixel)在HSI色彩空間所對應 的I值,毎值不同I值的像素數目;再將不同之I值由小 I 到大累加各值I值之像素數目,直至_累加到U值之像素 數目後超隹目之—半,則該I值邸為中值;及 ------- --- 該暫存器21則用來儲存該中值。 例如第五圃所示,為一 4X4的矩陣,其内之數字則是 毎一像素所對應之I值,該矩陣之像素,其毎齷不同I值 之像素數目的分佈曲嫌如第六所示,而像素在每镳I值 内的累計分佈曲》如箄;t·所示。由第五〜t麵可以看到 矩陣之16镳像素數目中超遇一半(10儀)儋素的I值小於 或等於2,因此這矩陣的中值是為2。 當取平均值時,本發明再次將上式(4)(5)(6)分別修 正如下: Mean.I <- Mean of I 第10頁 本紙張尺度適用中國國家標準(CNS ) A4规格(210X297公釐) 請' 先 閲 讀 背 之 注 2 奢 訂 五、發明説明(θ 15 A7 B7 經濟部中央樣準局貝工消费合作社印装 20Median.I Median of I Enhanced_R = (R-Median_I) xContrast_Enhance «ent_Gain + Median_I .............. (4) Enhanced_G = (G-Median_I) x Contrast_Enhanceient_6ain + Median_I ........... (5) Enbanced_B = (B-Nedian_I) xContrast_Enhanceaent_6ain + Median_I. ................ (6) To calculate the median value of a color image, please refer to the fourth figure. The two calculation units 2 include a median calculator 20 and a register 21, and use the median calculator 20 to count the I values corresponding to pixels in the HSI color space, and the number of pixels with different I values. Then add the number of different I values from small I to large and accumulate the number of pixels of each value I value, until _ accumulated to the number of pixels of U value is more than half of the number, then the I value is the median value; and ----- ---- The register 21 is used to store the median value. For example, as shown in the fifth garden, it is a 4X4 matrix, and the numbers in it are the I values corresponding to one pixel. The distribution of the number of pixels in the matrix and the number of pixels with different I values is like the sixth one. T, and the cumulative distribution curve of pixels within each 镳 I value is shown as 箄; t ·. From the fifth to t planes, it can be seen that half of the 16 pixels in the matrix are more than half (10 meters), and the I value of the element is less than or equal to 2, so the median value of this matrix is 2. When the average value is taken, the present invention modifies the above formulas (4), (5), and (6) respectively again as follows: Mean.I <-Mean of I Page 10 This paper size applies the Chinese National Standard (CNS) A4 specification ( 210X297 mm) Please read the note 2 of the back first. V. The description of the invention (θ 15 A7 B7) Printed by the Central Bureau of Specimen Bureau of the Ministry of Economic Affairs.
Enhanced.R=(R-Hean_I)x Contrast,Enhance*ent_Gain + Mean_I.............................(7) Enhanced_G=(G-Mean_I)x Contrast.Enhanceient.Gain +Mean_I.............................(8) Enhanced_B=(B-Mean_I)xContrast_Enhanceient_Gain + Hean_I.............................(9) 而要計篡一張彩f的平均值^讅參颺第八所示,該 第二計算簞元2包括一平均值計算器22及一暂存器23,而 以該平均值計算器22將所有像素之I值累加除以所有像索 數目,以输出所有像素之平均I值;及該暫存器23則用來 鍺存上述之平均I值。 提得一提的是,當上述所有像素數目設定為2的專次 方嫡像素,則在執行除法時就不需要除法器,只要把I的 累加值往左方移幾鶬位元邸可完成除法,如此節省硬讎上 之成本者。 同理,以第五鼸之矩陣為例,其算術平均值為2.12, 和實際中值有差距但不失為一可行的替代方法。 又,因参的對從《黼設計者的 _黏來看,計先為了统計像素I值 分佈的情形,就必須有足夠多的暫存器來記錄有多少籲倕 素落在逭些值I值上;再者,毎鶴暫存器所能«存 在最棰纗的情況下必須等於整張彩像像素的數目,所以毎 籲暫存器都必須夠大,更不用説累稹像素數目的累加器了 。比如轵一張352 X 240彩色的彩像就霈要25S®暫存器來 第11頁 本紙張尺度速用中國國家梯準(CNS ) A4規格(210X297公釐) Γ請先閲讀背面之注意事項再填寫本頁)Enhanced.R = (R-Hean_I) x Contrast, Enhance * ent_Gain + Mean_I ............. (7) Enhanced_G = (G-Mean_I) x Contrast.Enhanceient.Gain + Mean_I ............. (8) Enhanced_B = (B -Mean_I) xContrast_Enhanceient_Gain + Hean_I .............. (9) And calculate the average value of a color f ^ As shown in the eighth embodiment, the second calculation unit 2 includes an average calculator 22 and a register 23, and the average value of all pixels is divided by the number of all pixels by the average calculator 22 To output the average I value of all pixels; and the register 23 is used to store the above average I value. It is worth mentioning that when all the above pixels are set to the second-order square pixels, a divider is not required when performing the division, as long as the accumulated value of I is shifted to the left by a few bits. Divide, so save the cost on the hard disk. Similarly, taking the matrix of the fifth unit as an example, its arithmetic mean is 2.12, which is a gap from the actual median, but it is a feasible alternative. In addition, from the point of view of the "Designer's Guide", in order to count the distribution of pixel I values, there must be enough registers to record how many pixels fall on these values I Value; moreover, the crane register must be equal to the number of pixels of the entire color image in the most existing case, so the call register must be large enough, let alone the number of accumulated pixels The accumulator is up. For example, a color image of 352 X 240 color does not require a 25S® register. Page 11 Quickly use the Chinese National Standard (CNS) A4 size (210X297 mm) on this paper. Γ Please read the notes on the back first (Fill in this page again)
A7 B7 15 五、發明説明(?) 記錄I值的分佈,毎健暫存器都必須有17籲位元以記錄可 的@ 了值均和w的情況,而累稹像素數 目的加法器也必須為17值位元。 而利用平均值的對比加強,在硬體計算平均值並非如 吾人在直覺上所想的那般容易。首先必須有梅夠大的累加 器以累計彩像所有像素的I值,接著要有铕除法器,其除 數及被除數都相當的大。以上述352X 240的彩色彩像(非 2的霉次方)為例,所有像素的I累加值(被除數)可能 大至需要25價位元表示,而除數則為17籲位元,所以要實 現道值除法器的代價也相當得离。 基於上述方法及霣路裝置所需之««成本太高,於是 本發明藉由取樣待定數目的像袤夾KLiUJn所有的儋素求 取平價。取樣黏數目在本較佳實施例中均設定為二的軍 _______- , 次方以籣化除法器之運算;且取樣參數(CTRFACT)則由影 像像素數目(Iiage_Size)、取樣酤數目(Saiple_No)、及 計數最大值(Max_Range)三键參數決定如下: 影像像素數目 計數最大值 (#*先閲讀背面之注意事項再填寫本頁} 經濟部中央橾隼局貝工消费合作社印装 20 彩嫌像素數目-取樣黏數目計數最大值-取樣參數 取樣參數=ceil 取樣黏數目X計數最大值 彩供儋素數目 (11) 第12頁 本纸張尺度逋用中國國家標準(CNS ) A4規格(210X297公釐) 經濟部中央橾準扃真工消费合作社印裝 A7 B7 五、發明説明(丨ο ) 1 其中彩像像素數目是由被處理之彩像大小而定,取樣 貼數目是指被取樣來計算平均值的儋素數目,且該取樣黏 數目和影像像素數目之比值卽為取樣率。 4 ceil 〃表示取最小整數,故當取樣確定後,依照上 5 式(Π),吾人可藉由讖整計數最大值使取樣參數符合硬體 上設計的要求,及取樣率的精確度。例如吾人認為合理的 取樣率(取樣點數目/所有像素數目)是在2%〜20%, 此時若計數最大值小於5,則依上式(11),其取樣參數皆 為〇,根本不可能進行取樣。若取樣參數要是大於零之整 1〇 數,則計數最大值必須大於50。從另一餾角度來看,若計 數最大值為256 ,同理,依上式(11),則無法B分2.39% 舆2.0%的取樣率,亦邸精確度不高,故計數最大值之上 限則可依所需要之精確度來設定。 藉由上述原理,本發明更包括一取樣運算單元4,請 15 參閲第九圓所示,以取樣上述解fi缩後彩色彩儋之部份像 素賫料至上述第一計算單元1。該取櫬蓮算單元4包括有 一第三計算單元40、一比較器41、一計數器42、及一取樣 黏數比較器43,其中: 該第三計算單元40包括有一累加器400,用以毎讀取一 20 嫡像素即累加上述設定之取樣參數(大於零之整數),而 输出一累加值,及一減法器401。 · 該比較器41,其參考值則為上述設定之計數最大值, 且當比較某»儋索之累加值大於該參考值時,即輸出一取 樣倍號至上述第一計算單元1及上述第三計算單元40,該 第13頁 本纸張尺度適用中國國家標準(CNS ) Α4規格(210X297公釐) (請先閲讀背面之注意事項再填寫本頁) 裝· ,ιτ 經濟部中央標準局負工消費合作社印裝 A7 B7_ 五、發明説明(") 1 第一計算單元1即計算出該點像素之對應I值,而該第三 計算單元40之減法器401即将該累加值減去參考值,並自 差值繼鑛累加取樣參數。 該計數器42,毎接受一次比較器4 2輪出之取樣倍號, 5 即累加>1",且輸出一累加值,其累加值邸為取樣點數 目。 該取樣點數比較器43,其參考值則為上述設定之取攆 黏數目,當上述計數器42输出之累加值等於取樣黏數目時 二 即輪出一結束取櫬倍號至上述第一計算單元1以停止取懞 10 黏之蓮算,且由該第一計算單元1将所有取樣點之對應I 值输入上述第二計算單元2,以输出一新I值者。 再請參閲第十圏所示,本發明將上述取樣麵算單元4 配合第四圓之第一計算單元1,並以求^取平均值的方式為 一較佳組合實施例,如上所述,本發明毎讀取一籲像素, 15 該取櫬運算單元4之比較器41邸輪出一取嫌倍號,及取樣 黏數比較器43输出一結束取樣倍號;此時,第一計算單元 1之計數器12毎接收到上逑比較器41輸出之取樣信號即累 加"1",該計數器12為一毎次數到2又會再從0開始數 起之計數器,其输出> 〇〃、* 1〃、或> 2〃作為選擇控 20 制倍號至多工器13以邐擇像素的R、G、或B至累加器11進 行累加到該累加器11接受上述取樣黏數比較器43輪出之结 束取櫬倍號為止。 接著,該累加器11邸將I的累加值输出至第二計算單 元2,執行将I的累加值除以取樣黏數目的蓮算,同理, __第14頁_ 本纸張尺度逍用中國國家標準(CNS ) ΑΊ規格(210X297公釐) I.-1------------裝------訂------\ i' (· 請先閲讀背面之注意事項再填寫本頁) 經濟部中央標準局男工消費合作社印製 A7 _B7_ 五、發明説明(/2) 1 值得一提的是當取樣黏數目設定為2的蓽次方時,則執行 除法時便不需要除法器蓮算,只要把I的累加值往左方移 幾值位元卽可完成除法,如此更可節省硬鼸上之成本。而 該第二計算單元2的輪出便是目前彩像畫面的I平均值, 5 供若是做靜態彩像的對比加強,則上述對比加強蓮算單元 3邸讀取原彩像之R、G、B及上述第二計算單元2之I平 均值做對比加強運算後再送往顧示器。在本較佳實施例中 ,該對比加強蓮算單元3如第十一_所示.包括有三fi差 值計算器30,分別接受上述像索資料之R、G、B值,及共 1〇 同接受上述I平均值,以分別输出該R、G、B輿新I值之 差值;三值對比加強增益蓮算器31,分別連接上述三僮差 值計算器30,以將上述差值放大;及三俪加法器32,分別 連接上述三值對比加強增益蓮算器31,及共同接受上述新 I值,以將放大後之差值再加回原減去之新I值,得到R’ 15 、G’、B’值,便是被對比加強後之像素者。 又,將上述的對比加強方法直接蓮用在動態影像時, 所遇到的困難就是影像的I平均值(邸新I值)的産生時 間較該張影像的願示時間有一張盡面的延遲,也就是説除 非有額外的暫存元件,否則當一張影像的平均I值被算出 2〇 的同時,該張彩像已願示完畢,所以該張彩像的對比加強 無法如靜態影像一般以本身的I平均值做蓮算。因此,在 做動態彩像的對比加強時,請參閲第十二園所示,本發明 在第二計算單元2内更包括一第二暫存器24及一權數乘加 單元25,其中該第二暫存器24接於該第二計算單元2之平 ___第15頁_ 本紙張尺度逍用中國國家標準(CNS ) A4規格(210X297公釐) -! —^1 - * - I III - I - 1 - H II . { n (t先閲讀背面之注意事項存填寫本頁) 經濟部中夬標準局貝工消費合作社印製 • A7 _______B7 ______ 五、發明説明(/3 ) 1 均值計算器22,以用來暫存前幾張影供的I平均值•然後 再經權數乘加單元25将逭些I平均值乘上濂當的權數後相 加後鍺存於第二計算單元2之暫存器23,以做為下一張彩 像對bb加強的I平均值,而上述前幾張影像,在硬體的變 5 化實施上,可經由調整決定I平均值是由先前2、4、8或 ---- 6張彩像的權數乘加者。 ....一. --" _是,本發明除了對比加強外並配合HSI色彩空間之 亮度I值的讕整,處理後的彩像在亮度及色彩飽和度上都 相當令人谋意,確實將解S縮後的彩像晝面棋期,色彩暗 10 淡之~«_黏解除,且其中之取樣方法,更使本發 ' — 一. 一 ' 此外,在上述本k明之較佳實施例中, 其解壓縮後之影像資料皆偽以RGB色彩空間做説明,而因 各龆不同的色彩空間僅為習知數值的轉換,是以,本發明 所用HSI色彩空間之亮度I值之中值或平均值亦可逋用於 15 YCbCr等色彩空鬨做對比加強者。 综上所述,本發明之彩色彩像對比加強之方法及其霣 路装置,確可藉由以上技術之掲露,連到預期之使用目的 ,以及功效之長足增進,且,本索於申請前並未有相同之 設計公醑使用,誡符發明專利之新潁性、實用性、進步性 20 及産業上利用價值等諸要件,爱依法提出専利申讅。 惟,以上所述者,僅為本發明之較佳實施例,大凡熟 悉此項技藉人士,依據本發明之精神所作之變化,仍應包 括於本案之專利鲔_内。 第16頁 ^紙張尺度適用中國國家梯準(CNS ) A4規格(210X297公釐) (請先閲讀背面之注意事項再填寫本頁) » n^i Ik.—· In . ,A7 B7 15 V. Description of the invention (?) To record the distribution of I values, the health register must have 17 bits to record the possible @ average and w values, and the adder that accumulates the number of pixels. Must be 17 value bits. However, using the average value to enhance the comparison, calculating the average value in hardware is not as easy as we intuitively think. First, there must be an accumulator large enough to accumulate the I values of all pixels of the color image, and then there must be a unitary divider, whose divisor and dividend are quite large. Taking the above-mentioned 352X 240 color image (not a power of 2) as an example, the I cumulative value (dividend) of all pixels may be as large as 25 valence units, and the divisor is 17 units, so The cost of implementing a track divider is quite significant. Based on the above method and the required cost of the road device, the cost is too high, so the present invention obtains a parity by sampling all the pixels of a predetermined number of image clips KLiUJn. In this preferred embodiment, the number of sampling sticks is set to two army _______-, the power of which is calculated by the unitized divider; and the sampling parameter (CTRFACT) is determined by the number of image pixels (Iiage_Size) and the number of sampling units (Saiple_No). ), And the maximum count (Max_Range) three-key parameters are determined as follows: Maximum number of image pixel counts (# * Read the notes on the back before filling this page} Printed by the Central Bureau of the Ministry of Economic Affairs Number of pixels-Maximum number of sampling sticks count-Sampling parameter = ceil Number of sampling sticks X counts maximum Number of color supply units (11) Page 12 This paper uses China National Standard (CNS) A4 (210X297) (Mm) Printed by the Central Government of the Ministry of Economic Affairs, Real Industrial Consumer Cooperatives A7 B7 V. Description of the Invention (丨 ο) 1 Where the number of color image pixels is determined by the size of the color image being processed, and the number of sample stickers refers to the sampled Calculate the average number of primes, and the ratio of the number of sampling sticks to the number of image pixels 卽 is the sampling rate. 4 ceil 〃 means taking the smallest integer, so when the sampling is determined, follow the formula 5 above. Π), we can round the maximum value to make the sampling parameters meet the design requirements on the hardware, and the accuracy of the sampling rate. For example, I think the reasonable sampling rate (the number of sampling points / the number of all pixels) is 2% ~ 20%, at this time if the maximum value of the count is less than 5, according to the above formula (11), the sampling parameters are all 0, it is impossible to sample at all. If the sampling parameter is greater than zero and the whole number is 10, the maximum value of the count is Must be greater than 50. From another perspective, if the maximum value of the count is 256, similarly, according to the above formula (11), the B-score cannot be 2.39% and the sampling rate is 2.0%, so the accuracy is not high, so The upper limit of the maximum counting value can be set according to the required accuracy. By the above principle, the present invention further includes a sampling operation unit 4, please refer to the ninth circle to sample the decolorized color color Some of the pixels are reported to the above-mentioned first calculation unit 1. The lotus calculation unit 4 includes a third calculation unit 40, a comparator 41, a counter 42, and a sampling viscosity comparator 43, wherein: The third calculation unit 40 includes an accumulator 400 for reading an 20 嫡 pixels accumulate the sampling parameters set above (an integer greater than zero), and output an accumulated value and a subtractor 401. · The comparator 41, whose reference value is the maximum count value set above, and when compared When the accumulative value of a certain cable is greater than the reference value, a sampling multiple is output to the above-mentioned first calculation unit 1 and the above-mentioned third calculation unit 40. The page 13 of this paper applies the Chinese National Standard (CNS) Α4 Specifications (210X297mm) (Please read the precautions on the back before filling out this page) Installation, ιτ Printed by the Central Standards Bureau of the Ministry of Economic Affairs, Consumer Cooperatives A7 B7_ 5. Description of the invention (1) First calculation unit 1 That is, the corresponding I value of the pixel at that point is calculated, and the subtractor 401 of the third calculation unit 40 subtracts the reference value from the accumulated value, and then accumulates the sampling parameters from the difference. The counter 42 receives the sampling multiples from the comparator 4 in two rounds, 5 is accumulated > 1 ", and outputs an accumulated value whose accumulated value is the number of sampling points. The reference value of the sampling point number comparator 43 is the number of sticking points set as described above. When the accumulated value output by the counter 42 is equal to the number of sampling sticks, two are rolled out and the number of times is taken to the first calculation unit. 1 to stop taking Mongolian 10 sticky lotus, and the first calculation unit 1 inputs the corresponding I values of all sampling points into the second calculation unit 2 above to output a new I value. Please refer to the tenth aspect, the present invention combines the above-mentioned sampling surface calculation unit 4 with the first calculation unit 1 of the fourth circle, and obtains the average value as a preferred combination embodiment, as described above. In the present invention, a pixel is read. 15 The comparator 41 of the fetching operation unit 4 outputs a fetching multiple, and the sampling sticky number comparator 43 outputs an end sampling multiple. At this time, the first calculation The counter 12 of the unit 1 receives the sampling signal output from the upper comparator 41 and accumulates "1". The counter 12 is a counter that counts from 2 to 2 and starts counting from 0 again. Its output is> 0. , * 1〃, or > 2〃 as a selection control 20 multiplying number to the multiplexer 13 to select the pixel R, G, or B to the accumulator 11 to accumulate to the accumulator 11 to accept the sampling sticky number comparator described above At the end of the 43 rounds, take the double. Then, the accumulator 11 outputs the accumulative value of I to the second calculation unit 2, and executes the calculation of dividing the accumulative value of I by the number of sampling stickies. Similarly, __page 14_ This paper scale is easy to use China National Standard (CNS) ΑΊ Specifications (210X297 mm) I.-1 ------------ Installation ------ Order ------ \ i '(· Please first (Please read the notes on the back and fill in this page) A7 _B7_ printed by the Male Workers Consumer Cooperatives of the Central Standards Bureau of the Ministry of Economic Affairs 5. Description of the Invention (/ 2) 1 It is worth mentioning that when the number of sampling sticks is set to the power of two, The divider is not needed to perform the division. As long as the accumulated value of I is shifted to the left by a few bits, the division can be completed, which can save the cost on the hard disk. The rotation of the second calculation unit 2 is the current I average value of the color image frame. 5 If the contrast of the static color image is strengthened, the contrast enhancement unit 3 will read the R and G of the original color image. , B, and the average value of I of the second calculation unit 2 are compared and enhanced, and then sent to the indicator. In this preferred embodiment, the contrast-enhancing lotus calculation unit 3 is shown in the eleventh embodiment. It includes three fi difference calculators 30, which respectively accept the R, G, and B values of the above-mentioned image data, and a total of 10. The same average value of I is accepted to output the difference values of the new I values of R, G, and B respectively; the three-value contrast enhanced gain calculator 31 is connected to the three-child difference calculator 30, respectively, to convert the difference Zoom in; and the triplet adder 32, which is connected to the above three-value contrast enhancement gain calculator 31, and jointly accepts the new I value to add the amplified difference back to the original subtracted new I value to obtain R '15, G ', B' values are the pixels that have been enhanced by contrast. In addition, when the above-mentioned contrast enhancement method is directly applied to a dynamic image, the difficulty encountered is that the generation time of the I average value (the new I value) of the image has a full delay compared to the wish time of the image. That is, unless there are additional temporary storage components, when the average I value of an image is calculated as 20, the color image is ready to be displayed, so the contrast enhancement of the color image cannot be as static image. Use the average value of I to do the lotus calculation. Therefore, when contrast enhancement of dynamic color images is performed, please refer to the twelfth circle. The present invention further includes a second register 24 and a weight multiplying and adding unit 25 in the second calculation unit 2. The second register 24 is connected to the level of the second calculation unit 2 _ page 15 _ The paper size is in accordance with the Chinese National Standard (CNS) A4 specification (210X297 mm)-! — ^ 1-*-I III-I-1-H II. {N (t read the notes on the back first and fill in this page) Printed by Shelley Consumer Cooperative, China Standards Bureau, Ministry of Economic Affairs • A7 _______B7 ______ V. Description of Invention (/ 3) 1 Mean The calculator 22 is used to temporarily store the average I values of the first few films. Then the weighted multiplication and addition unit 25 multiplies these average values by the appropriate weights and adds them to the second calculation unit. The register 23 of 2 is used as the I average value of the next color image to bb enhancement, and the first few images mentioned above can be adjusted to determine the I average value from the previous 2, 4, 8 or ---- 6 multipliers and weights of color images. .... 一.-&Quot; _Yes, in addition to the contrast enhancement and the correction of the brightness I value of the HSI color space, the processed color image is quite interesting in terms of brightness and color saturation. , Indeed, the color image after the decompression of the day and day, the color is dark 10 ~ __ sticky release, and the sampling method among them makes this hair '— one. One' In addition, in the comparison of the above k In the preferred embodiment, the decompressed image data are all described in RGB color space, and each color space is only a conversion of a conventional value. Therefore, the brightness I value of the HSI color space used in the present invention is The median or average value can also be used for contrast enhancement such as 15 YCbCr. In summary, the method for enhancing the contrast of color images of the present invention and the road device thereof can indeed be connected to the intended purpose of use and the significant improvement of efficacy through the exposure of the above technology. There has not been the same design public use before. The newness, practicability, and advancement of the commandment invention patent20 and the industrial use value of the various elements, love to file a profit claim in accordance with the law. However, the above are only the preferred embodiments of the present invention. Anyone who is familiar with this technology and borrows according to the spirit of the present invention should still be included in the patent of this case. Page 16 ^ Paper size is applicable to China National Standard (CNS) A4 (210X297 mm) (Please read the notes on the back before filling this page) »n ^ i Ik.— · In.,
*1T* 1T
Claims (1)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW85110136A TW391130B (en) | 1996-08-20 | 1996-08-20 | Method and apparatus for contrast enhancement of color images |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW85110136A TW391130B (en) | 1996-08-20 | 1996-08-20 | Method and apparatus for contrast enhancement of color images |
Publications (1)
Publication Number | Publication Date |
---|---|
TW391130B true TW391130B (en) | 2000-05-21 |
Family
ID=21625403
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW85110136A TW391130B (en) | 1996-08-20 | 1996-08-20 | Method and apparatus for contrast enhancement of color images |
Country Status (1)
Country | Link |
---|---|
TW (1) | TW391130B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102750673A (en) * | 2011-11-30 | 2012-10-24 | 新奥特(北京)视频技术有限公司 | Method for improving image contrast |
CN104182947B (en) * | 2014-09-10 | 2017-04-26 | 安科智慧城市技术(中国)有限公司 | Low-illumination image enhancement method and system |
-
1996
- 1996-08-20 TW TW85110136A patent/TW391130B/en not_active IP Right Cessation
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102750673A (en) * | 2011-11-30 | 2012-10-24 | 新奥特(北京)视频技术有限公司 | Method for improving image contrast |
CN104182947B (en) * | 2014-09-10 | 2017-04-26 | 安科智慧城市技术(中国)有限公司 | Low-illumination image enhancement method and system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US7283663B2 (en) | Interpolation of edge portions of a digital image | |
Keelan | Handbook of image quality: characterization and prediction | |
TW391129B (en) | Apparatus and method for compressing image data outputted from image semsor having bayer pattern | |
US6642962B1 (en) | Merged pipeline for color interpolation and edge enhancement of digital images | |
US7664338B2 (en) | Perspective transformation of two-dimensional images | |
KR960009610A (en) | How to provide image scaling filters and video scaling | |
US20080123997A1 (en) | Providing a desired resolution color image | |
JP5219165B2 (en) | Raster image scaling method and scaling circuit | |
KR20070108221A (en) | Adaptive color interpolation | |
JPH11505350A (en) | Image Noise Reduction System Using Wiener Deformation Filter in Pyramid Image Display | |
CN108020519A (en) | A kind of virtual multiple light courcess spectrum reconstruction method based on color constancy | |
TW530485B (en) | Horizontal contour signal generating circuit for single board type color camera | |
US7751642B1 (en) | Methods and devices for image processing, image capturing and image downscaling | |
TW200824424A (en) | Adjusting apparatus for enhancing the contrast of image and method thereof | |
TW472211B (en) | Method and apparatus for interpolation | |
TW391130B (en) | Method and apparatus for contrast enhancement of color images | |
Andreadis et al. | Digital image scaling | |
JP2002278500A5 (en) | ||
EP2085926B1 (en) | Method and apparatus for generating thumbnail image | |
CN101370154A (en) | Image quality evaluating method and device | |
Nnolim | Design and implementation of novel, fast, pipelined HSI2RGB and log-hybrid RGB2HSI colour converter architectures for image enhancement | |
Murakami et al. | Piecewise Wiener estimation for reconstruction of spectral reflectance image by multipoint spectral measurements | |
Drew et al. | Multispectral processing without spectra | |
WO2023016365A1 (en) | Image processing method and apparatus, and electronic device | |
US20120148152A1 (en) | Image processing apparatus, image processing method, computer readable medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
GD4A | Issue of patent certificate for granted invention patent | ||
MM4A | Annulment or lapse of patent due to non-payment of fees |