DLCV 04 · A sharp guess can be less faithful
AdvancedDLCV 09 · Deep Image Restoration: Low-Level VisionInteractive plot
Description
A blurry observation fits two sharp clean signals. The blurry average has the lowest expected squared error; a sharp guess looks better but is usually farther from the truth.
Try this: Start from the sharp copy of A (commitment 1): its risk is higher than the blurry mean's. Slide the commitment to 0 and watch the risk fall to the minimum. Then raise the probability of A to 0.9: does the sharp copy of A catch up, and can it ever win?
Hints
The chapter section that uses this lab:
Explanation
Answer the prediction first, or open the explanation when you are ready.
Show the explanation
The mean has risk p(1−p) = 0.09 and the sharp copy of A has risk 1−p = 0.10. The gap shrinks as p approaches 1 but never closes, because the mean is the exact minimizer of expected squared error.
Lab · DLCV 04 · A sharp guess can be less faithful
- Expected reconstruction MSE
- 0.5
- Conditional-mean minimum MSE
- 0.25
- Sharpness (adjacent total variation)
- 3
Averaging incompatible possibilities minimizes expected squared error, but can remove structures present in every valid clean outcome. Choosing one restores structure while risking the wrong structure.
Try this: Start from the sharp copy of A (commitment 1): its risk is higher than the blurry mean's. Slide the commitment to 0 and watch the risk fall to the minimum. Then raise the probability of A to 0.9: does the sharp copy of A catch up, and can it ever win?
Two fixed 8-sample patterns, A = 00110011 and B = 11001100, are the only clean signals that could have produced the same blurry observation. P(A|y) = p is supplied, not learned. The blurry mean is p·A + (1−p)·B. Your reconstruction blends the mean toward A: (1−m)·mean + m·A, so m = 1 is exactly A (a sharp sample) and m = 0 is the mean. Risk = the expected mean squared error, p·MSE(A, output) + (1−p)·MSE(B, output), averaged over which pattern is the truth. Sharpness is the total variation (sum of absolute differences between neighbouring samples), not a perceptual quality score.