Separate Brightness from Color
A color leaf image needs a brightness change without shifting hue.
Read the full lesson and open its transcript below. A narrated video will be added after production.
What you will learn
A color leaf image needs a brightness change without shifting hue.
A practical question
A satellite team wants to brighten a leaf image so the veins are visible, but the green hue shifts. Instead of changing each red, green, and blue channel independently, consider a representation that separates a brightness-related component from color-related components.
The useful idea
That separation can make an intended brightness adjustment easier to describe. It isn't a guarantee of perfect color preservation. Converting between representations and changing one component can still affect saturation or values outside the display range. The exact meaning of a brightness channel depends on the color model we chose.
Work through the example
Place a known gray patch and a known green patch in the comparison. Change the brightness-related component, convert back, and inspect both patches. Did the gray remain neutral? Did the leaf's color stay useful for its actual analysis? Compare numerical channel values as well as the screen image. A colored part may appear darker under changing light without becoming a different material. Adjusting a brightness-related channel can make comparison easier, but color representations do not perfectly isolate physical illumination. Put a known reference color beside the part and compare both before and after. If that reference shifts unexpectedly, the adjustment could distort a classification decision. State which representation you changed and preserve the raw sensor values for later verification. Now check the idea on a concrete example: Modify the brightness-related channel and compare known colors before and after. A possible trap is worth naming: Different color spaces use different meanings for brightness; conversion must be explicit. The answer follows from that evidence: Some color representations separate a luminance-like component from chromatic components.
Limits and next step
Keep the sensor data untouched and label the adjusted image as a viewing copy. Next, we'll ask when a displayed pixel value can be treated as an amount of light, because many image calculations silently assume that it can.
Worked example
Modify the brightness-related channel and compare known colors before and after.
Quick check: test your understanding
Try it yourself: Consider this case: Modify the brightness-related channel and compare known colors before and after. Explain how you would check the result. In your answer, address this limitation: Different color spaces use different meanings for brightness; conversion must be explicit.
Check the reasoning
Some color representations separate a luminance-like component from chromatic components. Different color spaces use different meanings for brightness; conversion must be explicit.
What is the main idea behind separate brightness from color?
Some color representations separate a luminance-like component from chromatic components.
How is it used in the example?
Modify the brightness-related channel and compare known colors before and after.
What limitation should be checked?
Different color spaces use different meanings for brightness; conversion must be explicit.
Further reading
Digital Image Processing, 4th edition — Rafael C. Gonzalez and Richard E. Woods.
Original ToolsInVision teaching examples. Editorial approach · Suggest a correction
Accessible referenceLesson 53 transcript301 words · Narration ready for the upcoming videoOpen transcript
A satellite team wants to brighten a leaf image so the veins are visible, but the green hue shifts. Instead of changing each red, green, and blue channel independently, consider a representation that separates a brightness-related component from color-related components.
That separation can make an intended brightness adjustment easier to describe. It isn't a guarantee of perfect color preservation. Converting between representations and changing one component can still affect saturation or values outside the display range. The exact meaning of a brightness channel depends on the color model we chose.
Place a known gray patch and a known green patch in the comparison. Change the brightness-related component, convert back, and inspect both patches. Did the gray remain neutral? Did the leaf's color stay useful for its actual analysis? Compare numerical channel values as well as the screen image. A colored part may appear darker under changing light without becoming a different material. Adjusting a brightness-related channel can make comparison easier, but color representations do not perfectly isolate physical illumination. Put a known reference color beside the part and compare both before and after. If that reference shifts unexpectedly, the adjustment could distort a classification decision. State which representation you changed and preserve the raw sensor values for later verification. Now check the idea on a concrete example: Modify the brightness-related channel and compare known colors before and after. A possible trap is worth naming: Different color spaces use different meanings for brightness; conversion must be explicit. The answer follows from that evidence: Some color representations separate a luminance-like component from chromatic components.
Keep the sensor data untouched and label the adjusted image as a viewing copy. Next, we'll ask when a displayed pixel value can be treated as an amount of light, because many image calculations silently assume that it can.