some ideas

[hyupuunos]

This node has become a core component of my workflow, which has led to several new feature requests. Would you kindly consider whether these might be feasible?
​Mask Editor Settings Retention: The settings in the mask editor (e.g., brush size, brush color) appear to reset to default values even after adjustment. For example, I prefer using a white brush but have to reselect it every time. Could we implement a way to save these preferences between sessions?
​Reference Image Input for Image Filter: Would it be possible to add an input port to the image filter for receiving a reference image? This would be particularly useful for tasks like face-swapping, where seeing the original face as a reference during selection would improve accuracy.
​New "Image Choose" Node Proposal: Could we introduce a dedicated node (e.g., named "Image Choose") with multiple input ports to manually select between different images?
Use Case: For instance, generating 3 images from the same prompt using different checkpoints, then routing all outputs to this node for final selection.
​Batch Selection in Image Filter: Would supporting multi-image selection in the image filter for batch output be feasible? For example, selecting 3 out of 9 generated images for further processing.

[chrisgoringe]

You can already do most of this (unless I have misunderstood), because Image Filter can receive a batch of images, and you can select multiple images (you need to make sure the the global option "Send on Select" is turned off if you want to select more than one).

This also means you can do the "Image Choose" node - just use the built in Comfy 'Batch Images" node to combine them and feed them in. This is my main use case!

You could also use this to give you the reference image.

I'm not sure about the mask editor, that's a comfy component. I might take a look!

[hyupuunos]

okok,thanks~

I think I’m following your perspective now — I’ll take some time to review the details and reassess my approach.