Home

AI Photo Reimagination Masterclass

The P.A.S.S. Workflow

Preserve. Atmosphere. Surfaces. Subtleties.
A structured method to transform real photographs toward your intended vision — with physical realism, compositional integrity, and professional discipline.

Diagnose Direct Develop Debug Deliver
P A S S

The Complete Method

D.V.R.E. + P.A.S.S. = Full Workflow

Five stages from diagnosis to delivery. The DVRE framework reads the image; P.A.S.S. directs the transformation.

1

Diagnose DVRE

Use visual and structural analysis to understand the current image before changing anything.

Optical Analysis

  • Light direction (soft, hard, directional, diffused)
  • Shadow logic and existing exposure mood
  • Focal behavior (wide vs. tele compression)
  • Depth-of-field behavior

Structural Analysis

  • Primary anchors and leading lines
  • Spatial layers: foreground / middle / background
  • Material-critical surfaces (water, rock, snow, foliage)

Diagnostic Analysis

  • What is visually missing?
  • Which change creates highest impact?
  • Divergence lane recommendation (A/B/C)
2

Direct P.A.S.S.

Turn diagnosis into a precise, controllable edit plan using the four P.A.S.S. pillars.

P

Preserve

What must NOT change.

A

Atmosphere

The macro condition shift.

S

Surfaces

How materials react physically.

S

Subtleties

Micro-details that sell realism.

P.A.S.S.+ Extended Controls

Optics
Preserve perspective / compression
Color Discipline
Avoid hyper-real saturation
Post Character
Photographic tonal roll-off
Exclusions
Lock structures, faces, subjects
3

Develop

Apply transformations in a controlled multi-turn sequence. Don't try to solve everything in one pass.

3a
Initial AI Transformation

Upload photo + structured P.A.S.S. prompt to your chosen AI platform.

3b
Iterative Refinement

Multi-turn conversation to dial in weather, light, and atmosphere.

3c
Selective Inpainting

Target specific regions without re-processing the entire image.

4

Debug Realism Audit

Audit for realism failures, artifact patterns, optical mismatch, and stylistic overreach.

Light Logic

  • Shadow direction matches stated light source
  • Reflections agree with sky and sun angle

Material Physics

  • Snow accumulates where it should
  • Wet surfaces darken correctly

Botanical Accuracy

  • Conifers stay evergreen; deciduous trees change for season

Optical Integrity

  • Depth of field consistent across edited zones
  • Grain / detail character unified

Pattern Detection

  • No repeated tiles or AI blur halos

Scene Identity

  • Location still feels like itself; improved, not dramatized
The 5-Second Test: If a visually literate viewer cannot quickly detect artificiality, the edit is likely successful.
5

Deliver

Finalize, disclose appropriately, archive prompts, and build repeatable workflows.

Export at highest practical quality
Reintroduce natural texture or grain
Match sharpness across original and edited zones
Balance contrast and white balance globally
Archive prompt, rationale, lane choice, and disclosure note

Structured Divergence

Three Transformation Lanes

Not every edit should have the same divergence level. Choose a lane before you start.

Lane A

Purist

Correct or enhance only what plausibly could have happened with slightly better conditions.

Lane B

Storyteller

Make controlled environmental changes that alter mood or season while preserving scene truth.

Lane C

Visionary

Higher-divergence creative reimagination, labeled clearly as interpretive art.

Signature Case Studies

01Weather Change

Panlong Ancient Road (盤龍古道)

Xinjiang, China

Transformation: arid desert serpentine road → heavy snowfall with snow-covered mountains, road with snow layer and tire tracks, falling snow particles.

Key Insight: preserving the distinctive blue guardrails as a color anchor against the monochrome snow.

02Weather + Ice

High-Altitude Lake (班公湖 Style)

Pangong Tso, Ladakh

Transformation: HDR turquoise lake with boardwalk → active snowfall, partially frozen lake with translucent ice, cleaner blue sky, snow on shoreline.

Key Insight: "partially frozen — translucent ice with original turquoise visible beneath" maintains the lake's identity.

03Season Shift

Blausee / Blue Lake (藍湖)

Switzerland

Transformation: summer emerald lake with rowboat → peak autumn — deciduous trees blazing gold/orange/red, conifers remain green, fallen leaves on water.

Key Insight: "convert deciduous trees ONLY, keep conifers/evergreens dark green" — botanical specificity makes or breaks autumn realism.

Platform-Specific Workflows

AI Tool Ecosystem

Two primary platforms plus complementary finishing tools, each with distinct strengths.

Google Gemini

Gemini 2.5 Flash / 3 Pro Image

  • Upload photo directly in chat
  • Natural language transformation
  • Multi-turn editing conversations
  • Gem creation for repeatable workflows
Strength: excellent at preserving original composition.

ChatGPT

GPT-4o image generation

  • Upload + describe transformation
  • Iterate with follow-up instructions
  • Up to 10 images per batch
  • Custom GPT for studio workflows
Strength: superior photorealistic detail in complex scenes.

Complementary Tools

Specialized integrations

  • Luminar Neo: AI sky replacement + atmosphere
  • Lightroom / Camera Raw: AI landscape masking
  • Topaz Labs: sharpening + denoise
  • Photoshop: local cleanup + compositing
Role: precision finishing after AI generation.

Precision Integration

Masking, Inpainting & Blending

Move beyond whole-image prompting. The most convincing professional edits are selective.

Masking Strategy

Isolate regions for targeted control. AI handles the concept; you define the boundaries.

  • Region-based prompting for specific zones
  • Protect geological truth while changing atmosphere
  • Edge discipline between edited and original areas

Inpainting Logic

Add missing environmental elements without disrupting anchored geometry.

  • Add waterfall flow to dry channel (keep rock structure)
  • Enhance sky only while matching ground light
  • Freeze lake edges but not open central water

Compositing Refinement

Photoshop / Lightroom layer work to seamlessly marry AI output with original integrity.

  • Layer blending modes for natural transitions
  • Luminosity masking for selective adjustments
  • Frequency separation for texture matching

Professional Polish

Photoshop & Lightroom Finishing

AI generates the vision. Traditional tools deliver gallery-grade, client-safe, print-ready results.

Lightroom / Camera Raw

Global tonal control & color grading

1
White Balance Unification

Ensure the entire image lives in one believable color temperature.

2
AI Landscape Masking

Sky / Subject / Trees masks for targeted adjustments.

3
Tone Curve Refinement

Restore photographic tonal roll-off; AI outputs tend to be linear.

4
Saturation Control

Pull back any hyper-real colors with the HSL panel.

Photoshop

Local precision & detail work

1
Halo & Edge Correction

Clone stamp + healing brush at 100% to fix transition halos.

2
Texture & Grain Matching

Match AI-generated areas to the original sensor noise profile.

3
Sharpness Harmonization

Match sharpness between AI-edited and original zones.

4
Detail Painting

Hand-paint micro reflections, edge frost, water droplets.

Topaz Labs

Output enhancement

1
Denoise AI

Clean any AI artifacts while preserving legitimate texture detail.

2
Sharpen AI

Final output sharpening calibrated for delivery medium.

Ethical Disclosure

When sharing AI-enhanced photos, disclosure level scales with context.

Fine Art

Disclosure recommended; artistic license broadly accepted.

Editorial / Contest

Mandatory disclosure — many competitions require it.

Commercial / Client

Full transparency about extent of AI transformation.

Journalism

AI environmental transformation generally not acceptable.

Template: "Original photograph captured on [date] at [location]. Atmospheric/seasonal elements enhanced using AI tools to realize the photographer's creative vision. Core composition and subjects are unaltered."

Try It Now

Build a P.A.S.S. Prompt

Sign up for 3 free generations. No credit card required.

Launch Prompt Generator →
返回頂端