Depth Anything V3

DepthAnything_V3

Unified Depth Anything V3 node - all outputs, multiple normalization modes. **Normalization Modes:** - Standard: Original V3 min-max normalization (0-1 range, includes sky) - V2-Style: Disparity-based with content-aware contrast (default, best for ControlNet) - Sky appears BLACK (like V2) - Content-only normalization with percentile-based contrast - Enhanced depth gradations via contrast boost - Subtle edge anti-aliasing for natural transitions - Contribution by Ltamann (TBG) - Raw: No normalization, outputs metric depth (for 3D reconstruction/point clouds) **Outputs (always available):** - depth: Depth map (normalized or raw, depending on mode) - confidence: Confidence map (normalized 0-1) - ray_origin: Ray origin maps (for 3D, normalized for visualization) - ray_direction: Ray direction maps (for 3D, normalized for visualization) - extrinsics: Camera extrinsics (predicted camera pose) - intrinsics: Camera intrinsics (predicted camera parameters) - sky_mask: Sky segmentation (1=sky, 0=non-sky, Mono/Metric models only) - gaussian_ply_path: Path to raw 3D Gaussians PLY (Giant model only, empty string if not supported) **Optional Inputs:** - camera_params: Connect DA3_CreateCameraParams for camera-conditioned estimation - resize_method: How to handle patch size alignment (resize/crop/pad) - invert_depth: Toggle output convention. OFF (default): close=bright. ON: far=bright. - keep_model_size: Keep model's native output size instead of resizing back **Note:** Ray maps and camera parameters only available for main series models. Sky mask only available for Mono/Metric/Nested models. Connect only the outputs you need - unused outputs are simply ignored.

Pack: ComfyUI-DepthAnythingV3

custom_nodes.ComfyUI-DepthAnythingV3