This is an extension node for ComfyUI that allows you to load frames from a video in bulk and perform masking and sketching on each frame through a GUI.
Add this repository to the custom_nodes/ directory.
This is in the Image category.
- sequence_id: Please ignore (used for internal processing only).
- upload button: Specify the directory containing frames from the video using the dialog. (Use ffmpeg or similar to extract frames from the video.)
- start_index: Specify the start frame number. 0 to disable.
- end_index: Specify the end frame number. 0 to disable.
- images: Loaded frame data. If sketching is applied, it will be reflected in this output.
- mask_images: Masks for each frame are output as images. You may need to convert them to mask data using a Mask To Image node, for example.
- image_count: Number of processed frames.
Instead of specifying a directory containing frames, you specify a video file. Currently, only mp4 format is supported.
Otherwise, it is the same as the SequentialImageLoader Node.
(I used getVideoFrames.js for extracting frames from MP4)
- After loading the frames, right-click the node and select "Open In MaskEditor".
- In the editor that appears, perform masking and sketching as needed. It is based on the standard Mask Editor.
- Switch between automatic masking (magic wand) and manual masking.
- Switch between masking and sketching.
- Canvas. Right-click to mask, left-click to unmask.
- Change edit frame.
- Clear mask on the current frame.
- Paste the mask from the previous frame to the current frame.
- Change the thickness of the masking.
- Undo/Redo operations. Shortcut keys are alt+Z/shift+alt+Z. ( not Ctrl+Z! That's the standard shortcuts for ComfyUI. It might cause unexpected behavior.)
- Save or cancel the results.
Workflow (It's depend on ComfyUI-AnimateDiff-Evolved)
Video source
Temporary frame data accumulates in the input/ directory, so please delete data that is no longer needed at an appropriate time.