This page visualizes environment map generation results organized by dataset → scene → image. Each image shows the lighting modalities (envmap, irradiance, etc.) used to condition the generation, followed by the generated results for different random seeds, and finally the ground truth target image with a red border.
| Row Type | Description | Visual Indicator |
|---|---|---|
| Controls | Lighting modalities (e.g., envmap, irradiance) used as input conditions for generation | Label: "Controls:" | Standard border |
| Generated (seed N) | Images generated using the lighting conditions with different random seeds. Multiple rows show different seed variations. | Label: "Generated (seed N):" | Standard border |
| Ground Truth | The target reference image that the model is trying to generate | Label: "Ground Truth:" | Red border |