The Process

---

Overview

• Convert both input images to LMS colorspace
• Convert both input images to lαβ colorspace
• Transfer image1 colors onto image2 lαβ colorspace
• Convert image2 back to LMS colorspace
• Convert image2 back to RGB colorspace

---

Programming Notes

• Floating point numbers 0.0 .. 1.0 were used for pixel values
• Math.log10 and Math.pow are functions used

---

Step 1 - Converting from RGB to LMS Colorspace

For both images, apply this matrix to each pixel

Then, for each pixel, and each channel L, M, S, let L=logL, M=logM, and S=logS

---

Step 2 - Converting from LMS to lαβ Colorspace

For both images, apply this matrix to each pixel

And then apply this matrix to the pixels of both images

Step 3 - Transferring image1's colors to image2

First, do these steps:

• For each image, find the average l, α, and β values
• For each image, find the standard deviation of each channel l, α, and β
• Calculate image1's standard deviation / image2's standard deviation for each channel

After finding these variables, complete the following mathematical steps on image2's lαβ data:

• For each channel, subtract image2's average values found previously
• For each channel, multiply the ratio of standard devations that correspond with the channel
• For each channle, add image1's average values found previously

---

Step 4 - Converting from lαβ to LMS Colorspace

Only for image2, apply this matrix to each pixel

And then apply this matrix to the pixels

---

Step 5 - Converting from LMS to RGB Colorspace

Then, for each pixel in image2, and each channel L, M, S, let L=L^10, M=M^10, and S=S^10. Then apply this matrix