MoGePanoramaInference - ComfyUI Built-in Node Documentation

This node performs depth estimation on equirectangular panorama images. It splits the panorama into 12 perspective views, runs the MoGe depth estimation model on each view, and merges the per-view results back into a single depth map covering the full panorama. Per-view predicted normals and metric scale are ignored, because per-view scales would not align across the overlap seams.

Inputs

Parameter	Description	Data Type	Required	Range
`moge_model`	The MoGe model to use for inference.	MOGE_MODEL	Yes
`image`	Equirectangular panorama (any aspect). The node accepts a single image only; passing a batch of images raises an error. Only the first 3 color channels (RGB) are used.	IMAGE	Yes
`resolution_level`	Per-view detail (0 = fastest, 9 = most detailed) (default: 9).	INT	Yes	0 to 9
`split_resolution`	Resolution of each perspective split (default: 512).	INT	Yes	256 to 1024
`merge_resolution`	Long-side resolution of the merged equirect distance map (default: 1920).	INT	Yes	256 to 8192
`batch_size`	Views per inference batch (12 splits total) (default: 4).	INT	Yes	1 to 12

Notes:

The input must be a single image. If image contains more than one image in the batch, the node raises an error.
merge_resolution is treated as a maximum long-side size. If the panorama is smaller than this value, the merged map is produced at the original panorama size instead of being upscaled. The merged result is resized back to the original panorama resolution before being returned.

Outputs

Output Name	Description	Data Type
`moge_geometry`	A dictionary containing the estimated geometry: `points` (3D point cloud), `depth` (depth map), `mask` (valid area mask), and `image` (the input image).	MOGE_GEOMETRY

This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub

Source fingerprint (SHA-256): d35b6d42a5bb17c184bc56fe3867d3a183017084dc81649c0663a9fba2362770

​Inputs

​Outputs

Inputs

Outputs