Model
Trends
.ai
Model
Trends
.ai
最好的开源 AI 模型

ComfyUI 中的 Load Image (as Mask) 节点

Load Image (as Mask) turns one channel of a picture (alpha, red, green or blue) into a MASK, so black-and-white mask files or painted layers can drive inpainting and compositing.
节点包:ComfyUI core
分类:遮罩与局部重绘
此页面尚未翻译,因此以英文显示。

Load Image (as Mask) 的作用

Where Load Image gives you the picture, Load Image (as Mask) gives you a mask built from a chosen channel. Pick alpha for transparent PNGs, or red for a plain black-and-white mask file: white pixels become 1 (affected), black pixels 0 (protected). The output is a single-channel MASK at the size of the file.
It is the standard way to bring in a mask drawn in another program, a mask exported from a segmentation model, or a soft gradient mask for blending. Combine it with Grow Mask and Feather Mask to soften the edge before inpainting.

输入

名称
类型
含义
image
COMBO/upload
The mask file from ComfyUI/input.
channel
COMBO
alpha, red, green or blue. The channel that becomes the mask.

输出

名称
类型
含义
MASK
MASK
A float mask with 1 where the channel is bright (or transparent for alpha).
AD
跳过安装:云端预装的 ComfyUI 工作流
BitVector 在自家 GPU 上运行现成的 ComfyUI 工作流。无需安装,没有缺失节点,没有红框。用手机或笔记本打开,一分钟出图。

如何使用 Load Image (as Mask)

  1. Export a black-and-white mask at the exact size of the picture you will inpaint.
  2. Add Load Image (as Mask), upload it and set channel to red.
  3. Optionally add Grow Mask (6-16 px) and Feather Mask for a soft edge.
  4. Connect MASK to Set Latent Noise Mask or VAE Encode (for Inpainting).
  5. Check the mask with a Convert Mask to Image -> Preview Image pair if the result looks wrong.

设置与技巧

  • For alpha masks remember that transparent means masked (1), the opposite of a white-on-black file.
  • Masks are resized to the latent automatically by the inpainting nodes, but a mismatched aspect ratio is stretched.
  • Grey values make partial masks; a 50% grey pixel is half-strength denoise.
  • Invert Mask flips the selection without re-exporting the file.

Load Image (as Mask) 故障排查

The mask selects the wrong area (inverted)

原因
White and black were swapped in the file, or alpha was chosen for an opaque image.

修复方法
Add Invert Mask, or switch channel to red for a black-and-white file.

Mask is all zeros

原因
The chosen channel is uniform: for example alpha on a JPEG, which has no transparency.

修复方法
Use the red channel for black-and-white files, or re-export the mask with real transparency.

Hard, visible seam at the edge of the inpainted region

原因
A binary mask with no soft edge and a high denoise value.

修复方法
Add Grow Mask then Feather Mask (8-24 px) before the inpainting node, and lower denoise slightly.

AD
在手机上运行这个工作流
本页的每个工作流都已预装在 BitVector 上,模型和自定义节点一应俱全。选一个,输入提示词,完成。零配置,无需下载。

关于 Load Image (as Mask) 的问题

Can I paint the mask instead of loading a file?

Yes: use Load Image and its Mask Editor. Load Image (as Mask) is for masks that already exist as files.

Does the mask need the same size as the picture?

It should. Nodes resize it to fit, but different aspect ratios distort the selection.

相关节点

Load Image
ComfyUI core
Load Image reads a picture from the ComfyUI input folder (or an upload) and outputs an IMAGE tensor plus a MASK taken from the alpha channel, the starting point for img2img, inpainting, ControlNet and IP-Adapter graphs.
Grow Mask, Feather Mask and Invert Mask
ComfyUI core
The core mask tools expand (Grow Mask), soften (Feather Mask) and flip (Invert Mask) a MASK so inpainted regions blend into the surrounding picture without visible seams.
Set Latent Noise Mask
ComfyUI core
Set Latent Noise Mask attaches a MASK to a LATENT so the sampler only changes the white area: inpainting and regional regeneration with any normal checkpoint.
VAE Encode and VAE Encode (for Inpainting)
ComfyUI core
VAE Encode converts a pixel IMAGE into a LATENT for img2img; VAE Encode (for Inpainting) does the same while blanking the masked area so an inpainting model can repaint it.
AD
修节点修累了?交给云端
BitVector 在高速云 GPU 上维护着数百个 ComfyUI 工作流,已安装、已更新、已测试。没有 Python,没有 CUDA 报错,没有显存限制。任何浏览器都能用。

更多 ComfyUI 节点

CLIP Text Encode (Prompt)
ComfyUI core
CLIP Text Encode turns a text prompt into CONDITIONING using the model text encoder. One node holds the positive prompt, a second one the negative prompt.
ComfyUI Manager
ComfyUI-Manager
ComfyUI Manager is the extension that installs, updates and fixes custom node packs and models from inside the interface, resolves missing nodes in imported workflows and snapshots your setup.
KSampler
ComfyUI core
KSampler runs the denoising loop: it takes the model, prompts and a latent and produces the finished latent image, controlled by seed, steps, cfg, sampler, scheduler and denoise.
Load Checkpoint
ComfyUI core
Load Checkpoint (CheckpointLoaderSimple) opens a .safetensors or .ckpt model file and hands out the three parts every workflow needs: the diffusion MODEL, the CLIP text encoder and the VAE.
Save Image
ComfyUI core
Save Image writes the IMAGE tensor to ComfyUI/output as a PNG, with the whole workflow embedded in the file metadata so the picture can be dragged back into ComfyUI to restore the graph.
VAE Decode
ComfyUI core
VAE Decode converts the sampled LATENT into a pixel IMAGE with the VAE; VAE Decode (Tiled) does the same in tiles for very large images.