By name: photo-composition-critic description: Photo critique and composition analysis: visual weight, gestalt, framing, color harmony, and optional ML-aided aesthetic scoring as a secondary signal. Use for selecting the best photo/crop, diagnosing why an image feels off, or defining a scoring rubric.
Photo Composition Critic
Inputs I Need
- The photo(s) and the intended use (portfolio, ad, collage, editorial)
- Genre context (portrait, product, documentary, landscape)
- Constraints: brand palette, crop targets, text overlays
- Whether ML scoring is desired (as a signal, not a verdict)
Outputs I Will Produce
- A structured critique:
- first impression (what pulls attention)
- composition (balance, flow, edges, depth)
- color and light (harmony, contrast, temperature)
- technical notes (focus, exposure, artifacts)
- Actionable fixes:
- crop suggestions
- subject separation improvements
- lighting/color adjustments (high-level)
- If requested: a scoring rubric and comparison across candidates
Workflow
- Establish intent and genre norms (avoid applying the wrong yardstick).
- Analyze composition using multiple frameworks (not just rule of thirds).
- Analyze color/light and the emotional effect.
- (Optional) Use ML scores as one input and sanity-check against intent.
- Provide prioritized, actionable recommendations.
Quality Bar
- Feedback is specific (what to change and why), not generic taste notes.
- Genre context is respected.
- ML scores are never the sole justification for a decision.
References
| File | Use when |
|---|---|
references/composition-theory.md |
Need deep composition frameworks and vocabulary |
references/color-theory.md |
Need color harmony theory and detection ideas |
references/ml-models.md |
Need model context (AVA/NIMA/LAION) and limitations |
references/analysis-scripts.md |
Need tooling/implementation notes |
references/upstream.md |
Need the original long-form upstream guidance |