Tool: image analysis (artwork)

Image analysis - objective parameters

This page generates a fully browser-based, formal image analysis for an uploaded artwork (painting). After upload, it calculates tonal and contrast metrics (mean luminance, contrast spread, black/white clipping), color indicators (saturation, warm/cool share), a dominant color palette (K-means clusters with proportions), edge and line metrics (gradient energy and horizontal/vertical/diagonal orientation dominance), focus and visual-guidance data (top 10% contrast regions, centroid), proximity to the rule of thirds and golden ratio, symmetry and balance values (mass centroids and imbalance scores), as well as texture/brushstroke proxies and color-harmony heuristics. Results are output as text and color swatches.

Upload & analysis

Choose an image (e.g. a photo of a painting). The analysis runs entirely in the browser – no data is transmitted. The output consists of formal metrics commonly used as a basis for image discussion and analysis (no “AI interpretation”).

Results (text)

No image analyzed yet.

Dominant colors

Vorschau
Preview (local)

What is calculated?

All values are calculated from the pixels of the image, which may be downscaled internally. The metrics are formal: they describe brightness, color, structure, edges and spatial distribution – they provide no iconographic or semantic interpretation. Values are mostly output as proportions (0…1 or %) or as scores (0…1).

Tone & contrast

  • Mean brightness (luma) [0…1]
    Meaning: Average luminance (Rec.709) across all pixels.
    Interpretation: Low → overall dark/low-key; high → overall bright/high-key.
  • Contrast (std dev luma) [0…≈0.5]
    Meaning: Standard deviation of luma values.
    Interpretation: Higher → stronger light-dark tension; lower → flatter or more even tone values.
  • Clip black / clip white [0…1]
    Meaning: Share of pixels with ≤ ~2% luma or ≥ ~98% luma respectively.
    Interpretation: High values → loss of detail in shadows or highlights.

Color impression

  • Mean saturation [0…1]
    Meaning: Average HSV saturation.
    Interpretation: Low → desaturated or muted; high → color-intensive.
  • High / low saturation [0…1]
    Meaning: Share of pixels with saturation ≥ 0.6 or ≤ 0.2.
    Interpretation: Helps distinguish whether color accents occur only in spots or across larger areas.
  • Warm/cool share [0…1]
    Meaning: Hue heuristic (warm ≈ red/orange/yellow, cool ≈ blue). Neutral or desaturated areas may not count clearly.
    Interpretation: Warm dominates → a warmer impression; cool dominates → a cooler impression.

Dominant color palette

  • Top colors (K-means)
    Meaning: Cluster centers in RGB color space (typically 5–6 clusters) including the share of each cluster [0…1].
    Interpretation: Overview of the basic palette and weighting; large areas such as backgrounds may dominate.

Composition & lines

  • Edge/line energy [0…1]
    Meaning: Share of strong gradient pixels (Sobel) above a threshold, relative to the image.
    Interpretation: High → many edges/details/structure; low → flatter, softer, less edgy.
  • Orientation dominance [0…1]
    Meaning: Histogram of gradient angles (strong edges), summarized into horizontal/vertical/diagonal relative shares.
    Interpretation: Horizontal → calm or landscape-like impression, vertical → uprightness/architecture, diagonal → dynamics/movement.

Visual guidance / focus (top 10% contrast)

  • Focus area [0…1]
    Meaning: Share of pixels in the top 10% of local contrast values (|luma − local mean|).
    Interpretation: Small → few hard accents; large → contrast zones spread widely, producing a more restless focus.
  • Focus centroid (x,y) [0…1]×[0…1]
    Meaning: Normalized centroid of the focus pixels. (0,0)=top left, (1,1)=bottom right.
    Interpretation: Location of the visually strongest contrast center.
  • Rule-of-thirds score / golden-ratio score [0…1]
    Meaning: Proximity of the focus centroid to the respective intersection points (1 = very close).
    Interpretation: Higher → focus near classical composition anchors; lower → more central, marginal or unusual.

Symmetry & balance

  • Left/right and top/bottom symmetry [0…1]
    Meaning: Luma comparison of mirrored pixels (1 = very symmetrical).
    Interpretation: High → mirror-like order; low → asymmetric weighting.
  • Balance (visual mass)
    Meaning: “Mass” is weighted more strongly for darker pixels; calculated values include centroid (x,y), left/right and top/bottom imbalance [0…1], and corresponding scores [0…1].
    Interpretation: High imbalance → strong shift of visual weight; high center-balance score → mass close to the image center.

Texture / brushstroke proxies

  • Brushiness (Multi-Scale) [0…1]
    Meaning: Weighted combination of mean gradient energy across several scales (original, 1/2, 1/4).
    Interpretation: Higher → more grain, brush structure or fine variation; lower → smoother areas.
  • Laplacian variance (raw ≥ 0) & lapScore [0…1]
    Meaning: Laplacian variance as a sharpness/texture indicator; `lapScore` is a normalized heuristic mapping.
    Interpretation: Higher → stronger micro-contrast and edges; lower → softer, with less high-frequency detail.
  • Luma entropy [0…1]
    Meaning: Entropy of the tone-value histogram (64 bins), normalized.
    Interpretation: Higher → diverse distribution of tone values; lower → a few tone values dominate.

Color harmony (heuristics)

  • Analog / complementary / triadic [0…1]
    Meaning: Patterns are evaluated from a hue histogram (10° bins weighted by saturation × brightness): clusters (analog), opposites (≈180°), and triads (≈120°).
    Interpretation: Higher scores indicate the corresponding color relationships as a formal tendency, not as an aesthetic judgment.
  • Hue spread [0…1] & Concentration [0…1]
    Meaning: Entropy of the hue distribution (spread) and its inverse measure (concentration).
    Interpretation: High spread → many hue directions; high concentration → narrow, tonal palette.
  • Hue-Peaks (~°)
    Meaning: strongest hue bins (main directions).
    Interpretation: Rough indication of which color directions shape the image.