DIS Data Analytics 2 — Questions and Answers
Question 1: In digital imaging data analytics, what does a histogram primarily represent?
- The frequency distribution of pixel intensity values (Correct answer)
- The spatial arrangement of pixels in an image
- The color profile embedded in image metadata
- The compression ratio applied during encoding
Correct answer: The frequency distribution of pixel intensity values
A histogram plots the frequency of each pixel intensity value, revealing tonal distribution across shadows, midtones, and highlights.
Question 2: Which metric is most useful for quantifying image quality degradation introduced by lossy compression?
- Bits per pixel (BPP)
- Peak Signal-to-Noise Ratio (PSNR) (Correct answer)
- Color gamut coverage
- Dynamic range (stops)
Correct answer: Peak Signal-to-Noise Ratio (PSNR)
PSNR measures the ratio between maximum possible signal power and the power of corrupting noise, making it the standard lossy compression quality metric.
Question 3: A DIS analyst notices pixel values consistently clipped at 255 across a dataset. What does this indicate?
- Underexposure in the shadow regions
- Highlight overexposure causing blown-out areas (Correct answer)
- Gamma correction applied incorrectly
- Bit-depth mismatch during file conversion
Correct answer: Highlight overexposure causing blown-out areas
Pixel values clamped at 255 (the maximum for 8-bit) indicate highlight clipping, meaning detail in bright areas has been lost.
Question 4: What statistical measure best describes the spread of pixel intensity values in an image?
- Mean intensity
- Median intensity
- Standard deviation (Correct answer)
- Mode
Correct answer: Standard deviation
Standard deviation quantifies how widely pixel values deviate from the mean, reflecting the image's contrast and tonal range.
Question 5: When comparing two imaging systems, System A has lower bias but higher variance. System B has higher bias but lower variance. Which is preferred for consistent batch production?
- System A, because low bias ensures accuracy
- System B, because low variance ensures consistency (Correct answer)
- Neither, because both metrics must be minimized equally
- System A, because variance can be corrected in post
Correct answer: System B, because low variance ensures consistency
In production environments, low variance (consistency) is often preferred because systematic bias can be corrected through calibration, while high variance cannot.
Question 6: Which color model is most appropriate when analyzing perceptual uniformity in digital imaging quality control?
- RGB
- CMYK
- CIE Lab* (Correct answer)
- HSV
Correct answer: CIE Lab*
CIE Lab* is designed to approximate human vision's perceptual uniformity, making equal numerical differences correspond to equal perceived color differences.
Question 7: In a dataset of scanned documents, the signal-to-noise ratio (SNR) is measured at 20 dB. Doubling the SNR in linear terms corresponds to approximately how many dB increase?
- 3 dB
- 6 dB (Correct answer)
- 10 dB
- 20 dB
Correct answer: 6 dB
Because SNR in dB = 20·log10(linear ratio), doubling the linear SNR adds 20·log10(2) ≈ 6 dB.
DIS Data Analytics 2
In digital imaging data analytics, what does a histogram primarily represent?