1. Workflow Overview

This is an advanced lighting optimization workflow for photo editing and digital art, featuring:

• Intelligent lighting reconstruction with Xiaohongshu Cinematic Model

• Physically accurate lighting control via IC-Light

• IPAdapter+DetailTransfer dual-engine detail enhancement

• Adaptive Contrast-Aware Sharpening (CAS) + 4x upscaling

2. Core Models

3. Key Nodes

4. Workflow Stages

Phase 1: Input Preprocessing

• Input: Source image (e.g. 4181 (2).jpg)

• Processing:

  • Size normalization (1464px long edge)

  • VAE encoding to latent space

Phase 2: Lighting Optimization

• Core Techniques:

  • IC-Light (strength=0.345) for main light control

  • IPAdapter (weight=0.5) for feature retention

  • FreeU (b1=1.5, b2=1.6) for detail enhancement

Phase 3: Post-Processing

• DetailTransfer blends original/generated details

• CAS Sharpening (0.3) → 4x Upscale → Final output

5. Inputs & Outputs

Input Parameters:

• Image: Recommended ≥1464px (long edge)

• Prompt: "Optimize the lighting and details" (via easy positive node)

• Negative Prompt: Built-in hand/limb repair terms

Output:

• PNG with full generation metadata

6. Notes

1. VRAM: 12GB+ GPU recommended (for 1464px processing)

2. Critical Adjustments:

   • Light intensity: Modify ICLightConditioning's strength (default 0.345)

   • Detail retention: Adjust IPAdapterAdvanced's weight (default 0.5)

3. Troubleshooting:

   • Overexposure: Reduce FreeU_V2's backbone_factor

   • Over-sharpening: Decrease ImageCASharpening+ intensity