isoluminant colors
DESCRIPTION
Color2Gray: Salience-Preserving Color Removal Amy A. Gooch Sven C. Olsen Jack Tumblin Bruce Gooch. Visually important image features often disappear when color images are converted to grayscale . Isoluminant Colors. Color. Grayscale. Photoshop Grayscale. Traditional Methods. - PowerPoint PPT PresentationTRANSCRIPT
![Page 1: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/1.jpg)
Color2Gray: Salience-Preserving Color Removal
Amy A. Gooch Sven C. Olsen Jack Tumblin Bruce Gooch
![Page 2: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/2.jpg)
• Visually important image features often disappear when color images are converted to grayscale.
![Page 3: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/3.jpg)
Isoluminant Colors
Color Grayscale
![Page 4: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/4.jpg)
![Page 5: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/5.jpg)
Traditional Methods
• Contrast enhancement & Gamma Correction– Doesn’t help with isoluminant values
Photoshop Grayscale PSGray + Auto Contrast New Algorithm
![Page 6: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/6.jpg)
Simple Linear Mapping
Luminance Axis
![Page 7: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/7.jpg)
Principal Component Analysis (PCA)
Luminance Axis
![Page 8: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/8.jpg)
Problem with PCA
Worst case:Isoluminant Colorwheel
![Page 9: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/9.jpg)
Non-linear mapping
![Page 10: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/10.jpg)
• we want digital images to preserve a meaningful visual experience, even in grayscale.
![Page 11: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/11.jpg)
Goals
• Dimensionality Reduction – From tristimulus values to single channel
Loss of information
• Maintain salient features in color image– Human perception
![Page 12: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/12.jpg)
Original
Challenge:Many Color2Gray Solutions
.
.
.
![Page 13: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/13.jpg)
preserve the salient features of the color image
The algorithm introduced here reduces such losses by attempting to preserve the salient features of the color image. The Color2Gray algorithm is a 3-step process: 1. convert RGB inputs to a perceptually uniform CIE L a b ∗ ∗ ∗
color space, 2. use chrominance and luminance differences to create
grayscale target differences between nearby image pixels, and 3. solve an optimization problem designed to selectively
modulate the grayscale representation as a function of the chroma variation of the source image.
The Color2Gray results offer viewers salient information missing from previous grayscale image creation methods.
![Page 14: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/14.jpg)
Parameters
• θ: controls whether chromatic differences are mapped to increases or decreases in luminance value (Figure 5).
• α: determines how much chromatic variation is allowed to change the source luminance value (Figure 6).
• μ: sets the neighborhood size used for chrominance estimation and luminance gradients (Figure 7).
![Page 15: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/15.jpg)
![Page 16: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/16.jpg)
C2
C1
Map chromatic difference to increases or decreases in luminance values
Color Space
+b*
+a*-a*
![Page 17: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/17.jpg)
+DC1,2
q
vq
sign(||DCi,j||) = sign(DCi,j . vq )
Color Difference
Spacevq = (cos q, sin q)
+Db*
+Da*-Da*
-
-
+
+
-Db*
![Page 18: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/18.jpg)
Photoshop Grayscaleq = 225
q = 45
![Page 19: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/19.jpg)
Grayscale
q = 45q = 135
q = 0
![Page 20: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/20.jpg)
![Page 21: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/21.jpg)
How to CombineChrominance and Luminance
dij = (Luminance) DLij
![Page 22: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/22.jpg)
How to CombineChrominance and Luminance
dij =
DLij
||DCij||
(Luminance) if |DLij| > ||DCij||
(Chrominance)
![Page 23: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/23.jpg)
How to CombineChrominance and Luminance
d (a) ij =
DLij
crunch(||DCij||)
if |DLij| > crunch(||DCij||)
a
-a crunch(x) = a * tanh(x/a) -120, -120
120, 120
0
100
![Page 24: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/24.jpg)
.
.
.
How to CombineChrominance and Luminance
d(a,q)ij =
DLij
crunch(||DCij||) if DCij . nq ≥ 0
if |DLij| > crunch(||DCij||)
crunch(-||DCij||) otherwise
Grayscale
![Page 25: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/25.jpg)
Color2Grey Algorithm
Optimization:
min S S ( (gi - gj) - di,j )2 j=i-m
If dij == DL then ideal image is gOtherwise, selectively modulated by DCij
i
i+m
![Page 26: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/26.jpg)
![Page 27: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/27.jpg)
![Page 28: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/28.jpg)
Validate "Salience Preserving"
Apply Contrast Attention model by Ma and Zhang 2003
Original PhotoshopGrey Color2Grey
![Page 29: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/29.jpg)
Validate "Salience Preserving"
Original PhotoshopGrey Color2Grey
![Page 30: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/30.jpg)
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, VOL. 32, NO. 9, SEPTEMBER 2010
Color to Gray: Visual Cue PreservationMingli Song, Member, IEEE, Dacheng Tao, Member, IEEE, Chun Chen,Xuelong Li, Senior Member, IEEE, and Chang Wen Chen, Fellow, IEEE
• algorithm based on a probabilistic graphical model with the assumption that the image is defined over a Markov random field.
• color to gray procedure can be regarded as a labeling process to preserve the newly well-defined visual cues of a color image in the transformed gray-scale image
• three cues are defined namely, color spatial consistency, image structure information, and color channel perception priority
• visual cue preservation procedure based on a probabilistic graphical model and optimize the model based on an integral minimization problem.
![Page 31: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/31.jpg)
![Page 32: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/32.jpg)
![Page 33: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/33.jpg)
Colorization: History• Hand tinting
– http://en.wikipedia.org/wiki/Hand-colouring
![Page 34: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/34.jpg)
Colorization: History• Film colorization
– http://en.wikipedia.org/wiki/Film_colorization
Colorization in 1986 Colorization in 2004
![Page 35: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/35.jpg)
• Colorization by example
• Colorization using “scribbles”
![Page 36: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/36.jpg)
R.Irony, D.Cohen-Or, D.Lischinski
Eurographics Symposium on Rendering (2005)
Colorization by example
![Page 37: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/37.jpg)
MotivationColorization, is the process of adding color to monochrome images and video.
Colorization typically involves segmentation + tracking regions across frames - neither can be done reliably – user intervention required – expensive + time consuming
Colorization by example – no need for accurate segmentation/region tracking
![Page 38: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/38.jpg)
The method• Colorize a grayscale image based on a user
provided reference.Reference Image
![Page 39: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/39.jpg)
![Page 40: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/40.jpg)
Naïve MethodTransferring color to grayscale images [Walsh, Ashikhmin, Mueller 2002]
• Find a good match between a pixel and its neighborhood in a grayscale image and in a reference image.
![Page 41: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/41.jpg)
By Example Method
![Page 42: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/42.jpg)
Overview1. training2. classification3. color transfer
![Page 43: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/43.jpg)
Training stage
Input: 1. The luminance channel of the reference image2. The accompanying partial segmentation
Construct a low dimensional feature space in which it is easy to discriminate between pixels belonging to differently labeled regions, based on a small (grayscale) neighborhood around each pixel.
![Page 44: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/44.jpg)
Training stage
Create Feature Space(get DCT cefficients)
![Page 45: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/45.jpg)
Classification stage
For each grayscale image pixel, determine which region should be used as a color reference for this pixel.
One way: K-Nearest –Neighbor Rule
Better way: KNN in discriminating subspace.
![Page 46: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/46.jpg)
KNN in discriminating subspace
Originaly sample point has a majority of Magenta
![Page 47: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/47.jpg)
KNN in discriminating subspace
Rotate Axes in the direction of the intra-difference vector
![Page 48: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/48.jpg)
KNN in discriminating subspace
Project Points onto the axis of the inter-difference vector
Nearest neigbors are now cyan
![Page 49: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/49.jpg)
KNN Differences
Simple KNN
Discriminating KNN
Matching Classification
Use a median filter to create a cleaner classification
![Page 50: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/50.jpg)
Color transferUsing YUV color space
![Page 51: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/51.jpg)
Final Results
![Page 52: Isoluminant Colors](https://reader035.vdocument.in/reader035/viewer/2022081503/56816386550346895dd47064/html5/thumbnails/52.jpg)