ComfyUI-RMBG

AILab_ColorInput

Returns a validated hexadecimal color code, using either a named preset or a custom color value. Short hex forms are normalized before output.

🧪AILab/🛠️UTIL/🔄IO · 0 Eingaben · 2 Parameter · 1 Ausgabe

AILab_ColorToMask

Creates a mask by comparing image pixels with a selected color within a configurable threshold. The result can be inverted before output.

🧪AILab/🧽RMBG · 1 Ausgabe

AILab_CropObject

Finds the bounding box of the nonzero image or mask content, expands it by optional padding, and crops both the image and mask to that object region.

🧪AILab/🖼️IMAGE · 2 Eingaben · 1 Parameter · 2 Ausgaben

AILab_Florence2

Runs a selected Florence-2 model for text-prompted polygon segmentation, phrase grounding, or region proposals. It returns a rendered result image, selected or filled masks, and structured JSON containing the detected regions.

🧪AILab/🧽RMBG · 5 Eingaben · 4 Parameter · 3 Ausgaben

AILab_Florence2ToCoordinates

Converts Florence-2 bounding-box JSON into center-coordinate text, typed bounding boxes, and an optional raster mask aligned to a supplied image. It can select specific indices or gather boxes across the batch.

🧪AILab/🧽RMBG · 2 Eingaben · 2 Parameter · 3 Ausgaben

AILab_ICLoRAConcat

Concatenates an object image with an optional base image in left-right or top-bottom layout for IC-LoRA-style conditioning. It propagates or derives alpha masks and returns the combined image, both masks, output dimensions, and object placement coordinates.

🧪AILab/🖼️IMAGE · 4 Eingaben · 2 Parameter · 7 Ausgaben

AILab_ImageCombiner

Places a scaled foreground at a percentage position over a background and blends it with normal, multiply, screen, overlay, add, or subtract mode. Foreground opacity and an optional mask control the composite, which is returned with its dimensions.

🧪AILab/🖼️IMAGE · 3 Eingaben · 5 Parameter · 3 Ausgaben

AILab_ImageCompare

Builds a labeled comparison canvas from up to three images, using a selected image as the sizing reference and configurable label and background colors.

🧪AILab/🖼️IMAGE · 5 Eingaben · 4 Parameter · 1 Ausgabe

AILab_ImageCrop

Crops a pixel-sized region from an anchored position with x/y offsets and returns both the crop and a remainder output. Split mode assembles the image areas outside the crop; otherwise the remainder is a zero image.

🧪AILab/🖼️IMAGE · 1 Eingabe · 6 Parameter · 2 Ausgaben

AILab_ImageMaskConvert

Normalizes image and mask inputs into paired IMAGE and MASK outputs. It can expand a mask into an RGB image or derive a mask from an image’s alpha, red, green, or blue channel.

🧪AILab/🖼️IMAGE · 2 Eingaben · 1 Parameter · 2 Ausgaben

AILab_ImageMaskToList

Collects up to six image-and-mask pairs into parallel output lists, creating an empty mask when an image has none. Optional fit or center-crop modes align each pair to the first image’s dimensions.

🧪AILab/🖼️IMAGE · 12 Eingaben · 1 Parameter · 5 Ausgaben

AILab_ImagePreview

Displays an image batch in the ComfyUI preview UI while passing the original IMAGE through unchanged.

🧪AILab/🖼️IMAGE · 1 Eingabe · 0 Parameter · 1 Ausgabe

AILab_ImageResize

Resizes an image and optional mask by explicit dimensions, megapixels, scale factor, or longest/shortest-side target. It supports stretch, crop, colored or edge padding, blurred pillarboxing, device and interpolation choice, divisibility constraints, and returns final dimensions.

🧪AILab/🖼️IMAGE · 4 Ausgaben

AILab_ImageStitch

Stitches up to four images left, right, up, down, in a 2×2 grid, or in a Kontext layout. It can match sizes, add colored spacing, cap dimensions or megapixels, preserve batch/channel compatibility, and returns the composite dimensions.

🧪AILab/🖼️IMAGE · 5 Eingaben · 7 Parameter · 3 Ausgaben

AILab_ImageToList

Collects up to six image inputs into an IMAGE list and reports the first image’s width, height, and list size. Optional fit or center-crop modes align later images to the first image.

🧪AILab/🖼️IMAGE · 6 Eingaben · 1 Parameter · 4 Ausgaben

AILab_LamaRemover

Removes masked regions from images with the Big-LAMA inpainting model. Removal strength expands the effective area and edge smoothness softens the transition before inpainting.

🧪AILab/🧽RMBG · 2 Eingaben · 2 Parameter · 1 Ausgabe

AILab_LoadImage

Loads an uploaded image, local or network path, or URL and returns its pixels, alpha mask, width, and height. It can resize by megapixels, scale factor, width, height, or longest/shortest side with selectable resampling.

🧪AILab/🖼️IMAGE · 4 Ausgaben

AILab_LoadImageAdvanced

Loads an uploaded image, path, or URL with the pack’s resizing controls and derives a mask from alpha, red, green, or blue. It also returns an RGB mask preview, dimensions, and extracted image metadata.

🧪AILab/🖼️IMAGE · 6 Ausgaben

AILab_LoadImageBatch

Loads images from a directory or a comma/newline list of paths or URLs, with sequential, reverse, or random order plus start and count controls. Each image is resized independently and returned with parallel mask, width, and height lists.

🧪AILab/🖼️IMAGE · 4 Ausgaben

AILab_LoadImageSimple

Loads an uploaded image, local or network path, or URL without resizing and returns its pixels, alpha mask, width, and height.

🧪AILab/🖼️IMAGE · 0 Eingaben · 2 Parameter · 4 Ausgaben

AILab_MaskCombiner

Combines up to four masks after resizing them to the first mask’s shape. Union takes the per-pixel maximum, intersection takes the minimum of the first two, and difference returns their absolute difference.

🧪AILab/🖼️IMAGE · 4 Eingaben · 1 Parameter · 1 Ausgabe

AILab_MaskEnhancer

Refines a mask with sensitivity scaling, edge smoothing, hole filling, Gaussian blur, boundary expansion or contraction, and optional inversion.

🧪AILab/🖼️IMAGE · 1 Eingabe · 6 Parameter · 1 Ausgabe

AILab_MaskExtractor

Extracts or isolates the portion of an image selected by an optional mask. It can preserve transparency or composite the selected region over a chosen background color.

🧪AILab/🖼️IMAGE · 3 Eingaben · 2 Parameter · 1 Ausgabe

AILab_MaskOverlay

Creates a UI preview by overlaying a chosen color on an image through a mask at configurable opacity. It returns the composited preview image and original mask for downstream use.

🧪AILab/🖼️IMAGE · 3 Eingaben · 1 Parameter · 2 Ausgaben

AILab_MaskPreview

Renders a mask as a three-channel image in the ComfyUI preview UI while passing the original MASK through unchanged.

🧪AILab/🖼️IMAGE · 1 Eingabe · 0 Parameter · 1 Ausgabe

AILab_MaskToList

Collects up to six masks into a MASK list. Optional fit or center-crop modes align later masks to the first mask’s dimensions.

🧪AILab/🦠MASK · 6 Eingaben · 1 Parameter · 1 Ausgabe

AILab_Preview

Displays an optional image, mask, or both in the ComfyUI preview UI and passes the original values through separate IMAGE and MASK outputs.

🧪AILab/🖼️IMAGE · 2 Eingaben · 0 Parameter · 2 Ausgaben

AILab_ReferenceLatentMask

Prepares inpainting conditioning by attaching a latent image and processed mask to the conditioning metadata. It can expand or shrink and blur the mask, include the latent as reference conditioning, and optionally limit generation to the masked region.

🧪AILab/🎭Inpaint · 3 Eingaben · 3 Parameter · 3 Ausgaben

AILab_SDMatte

Extracts a high-quality foreground alpha matte with the selected SDMatte model, using a supplied mask or image alpha as guidance. It supports transparent-object handling, mask refinement and sensitivity, edge adjustments, device/resolution controls, and transparent or colored output.

🧪AILab/🧽RMBG · 3 Eingaben · 10 Parameter · 3 Ausgaben

AILab_UnbatchImages

Splits the first eight images of an input batch into eight separate IMAGE outputs. When the batch has fewer than eight images, the final available image is repeated to fill the remaining sockets.

🧪AILab/🖼️IMAGE · 1 Eingabe · 0 Parameter · 8 Ausgaben

AILab_YoloV8

Runs a selected local YOLOv8 model on image batches and returns annotated images, a merged detection mask, and individual masks. The simplified interface selects how many detections to merge and an optional starting mask index.

AILab_YoloV8Adv

Runs a selected local YOLOv8 model and returns annotated images, a merged mask, and individual detection masks. Advanced controls expose confidence, IoU, class filtering, device, detection limit, high-resolution masks, class-agnostic NMS, and mask selection.

🧪AILab/🧽RMBG · 3 Ausgaben

BiRefNetRMBG

Removes image backgrounds with a selected BiRefNet variant and returns the composited foreground, mask, and mask preview. It supports mask blur and boundary offset, inversion, foreground-color refinement, and transparent or solid-color backgrounds.

🧪AILab/🧽RMBG · 2 Eingaben · 7 Parameter · 3 Ausgaben

BodySegment

Segments selected human body classes—including hair, clothing, face, limbs, torso skin, glasses, and feet—with a human-parsing model. It returns a cutout image, mask, and mask preview with blur, offset, inversion, and background controls.

🧪AILab/🧽RMBG · 2 Eingaben · 4 Parameter · 3 Ausgaben

ClothesSegment

Segments selected clothing and person classes such as hats, upper clothes, dresses, pants, shoes, bags, hair, face, and limbs. It returns the composited image, mask, and mask preview with resolution and edge/background controls.

🧪AILab/🧽RMBG · 2 Eingaben · 5 Parameter · 3 Ausgaben

FaceSegment

Segments selected facial parts such as skin, eyes, brows, nose, lips, hair, ears, glasses, earrings, and neck. It returns the composited image, mask, and mask preview with resolution, blur, offset, inversion, and background controls.

🧪AILab/🧽RMBG · 2 Eingaben · 5 Parameter · 3 Ausgaben

FashionSegmentAccessories

Builds an accessory-selection value for Fashion Segment Clothing, covering wearable accessories and garment details such as hats, glasses, bags, belts, pockets, zippers, bows, and sequins.

🧪AILab/🧽RMBG · 0 Eingaben · 0 Parameter · 1 Ausgabe

FashionSegmentClothing

Segments selected garment categories and optional accessory/detail categories using a fashion-parsing model. It returns a composited image, mask, and mask preview with processing resolution, edge refinement, inversion, and background controls.

🧪AILab/🧽RMBG · 3 Eingaben · 5 Parameter · 3 Ausgaben

RMBG

Removes image backgrounds with a selectable RMBG-2.0, INSPYRENET, BEN, or BEN2 model. It returns the processed image, mask, and mask preview, with controls for sensitivity, mask blur and offset, inversion, foreground refinement, transparency, or a solid background color.

🧪AILab/🧽RMBG · 2 Eingaben · 8 Parameter · 3 Ausgaben

SAM2Segment

Uses GroundingDINO to find regions matching a text prompt and SAM 2 to segment those boxes. It returns a composited image, combined mask, and mask preview with model/device, threshold, edge, inversion, and background controls.

🧪AILab/🧽RMBG · 2 Eingaben · 9 Parameter · 3 Ausgaben

SAM3Segment

Segments text-prompted objects with SAM 3 and returns composited images, masks, and mask previews. It can merge detections or emit separate instances, rank and select segments, limit their count, adjust edges, unload the model, and choose transparent or colored backgrounds.

🧪AILab/🧽RMBG · 2 Eingaben · 11 Parameter · 3 Ausgaben

Segment

Uses GroundingDINO to detect boxes for a text prompt and the selected SAM model to create a combined segmentation mask. It returns the cutout or composited image, mask, and mask preview with threshold, edge, inversion, and background controls.

🧪AILab/🧽RMBG · 2 Eingaben · 8 Parameter · 3 Ausgaben

SegmentV2

Uses a selected GroundingDINO and SAM implementation to segment all regions matching a text prompt across an image batch. It returns the composited images, combined masks, and mask previews with detection, edge, inversion, and background controls.

🧪AILab/🧽RMBG · 2 Eingaben · 8 Parameter · 3 Ausgaben