bilateral reference framework (BiRefNet) for high-resolution dichotomous image segmentation (DIS)

Birefnet Background Removal

Detail

HTTPValidationError

Content Type

File Data

The name of the file. It will be auto-generated if not provided.

File Name

File Size

Height

The URL where the file can be downloaded from.

Width

Image

URL of the image to remove background from

Image Url


            Model to use for background removal.
            The 'General Use (Light)' model is the original model used in the BiRefNet repository.
            The 'General Use (Heavy)' model is a slower but more accurate model.
            The 'Portrait' model is a model trained specifically for portrait images.
            The 'General Use (Light)' model is recommended for most use cases.

            The corresponding models are as follows:
            - 'General Use (Light)': BiRefNet-DIS_ep580.pth
            - 'General Use (Heavy)': BiRefNet-massive-epoch_240.pth
            - 'Portrait': BiRefNet-portrait-TR_P3M_10k-epoch_120.pth
        

Model

The resolution to operate on. The higher the resolution, the more accurate the output will be for high res input images.

Operating Resolution

Output Format

Whether to output the mask used to remove the background

Output Mask

Whether to refine the foreground using the estimated mask

Refine Foreground

Input


            Model to use for background removal.
            The 'General Use (Light)' model is the original model used in the BiRefNet repository.
            The 'General Use (Light)' model is the original model used in the BiRefNet repository but trained with 2K images.
            The 'General Use (Heavy)' model is a slower but more accurate model.
            The 'Matting' model is a model trained specifically for matting images.
            The 'Portrait' model is a model trained specifically for portrait images.
            The 'General Use (Light)' model is recommended for most use cases.

            The corresponding models are as follows:
            - 'General Use (Light)': BiRefNet-DIS_ep580.pth
            - 'General Use (Heavy)': BiRefNet-massive-epoch_240.pth
            - 'Portrait': BiRefNet-portrait-TR_P3M_10k-epoch_120.pth
        

InputV2

Mask Image

Output

Location

Message

Error Type

ValidationError

FastAPI

Resize Image

Radius

BlurMaskInput

BlurMaskOutput

High threshold for the hysteresis procedure

High Threshold

Low threshold for the hysteresis procedure

Low Threshold

CannyInput

Color

Compare Text

Text to return if the input text does not match the compare text

Fail Text

Text to return if the input text matches the compare text

Return Text

Text

CompareTextInput

Background Image Url

Mask Image Url

Overlay Image Url

CompositeImageInput

Bbox

Det Score

Embedding File

Kps Image

FaceDetection

File

Pixels

The threshold to convert the image to a mask. 0-255.

Threshold

GrowMaskInput

GrowMaskOutput

ImageInput

ImageOutput

Image Size

ImageSizeOutput

Template

InsertTextInput

Det Size Height

Det Size Width

Max Face Num

Model Url

Sorting

Sync Mode

InsightfaceInput

Faces

InsightfaceOutput

InvertMaskOutput

MaskInput

Color to replace the transparent pixels with

Transparent Color

RGBAToRGBImageInput

Pattern

Replace

RegexReplaceInput

Position of cropping. Only used when mode is 'crop', default is center

Cropping Position

Mode

Color of padding. Only used when mode is 'pad', default is black

Padding Color

Resizing strategy. Only used when mode is 'scale', default is nearest

Resampling

Proportions of the image. Only used when mode is 'scale', default is fit

Scaling Proportions

ResizeImageInput

If set, the output dimensions will be divisible by this value.

Enforce Divisibility

Maximum number of pixels in the output image.

Max Pixels

ResizeToPixelsInput

ShrinkMaskInput

ShrinkMaskOutput

TeedInput

TeedOutput

TextOutput

The threshold to convert the image to a mask.

TransparentImageToMaskInput

TransparentImageToMaskOutput

Convert a Image with Alpha Channel to RGB with a background color

RGBA to RGB

Run Any Stable Diffusion model with customizable LoRA weights.

Stable Diffusion with LoRAs


            The scale of the control net weight. This is used to scale the control net weight
            before merging it with the base model.
        

Conditioning Scale

optional URL to the controlnet config.json file.

Config Url


            The percentage of the image to end applying the controlnet in terms of the total timesteps.
        

End Percentage

URL of the image to be used as the control net.


            The index of the IP adapter to be applied to the controlnet. This is only needed for InstantID ControlNets.
        

Ip Adapter Index


            The mask to use for the controlnet. When using a mask, the control image size and the mask size must be the same and divisible by 32.
        

Mask Url

URL or the path to the control net weights.

Path


            The percentage of the image to start applying the controlnet in terms of the total timesteps.
        

Start Percentage

The optional variant if a Hugging Face repo key is used.

Variant

ControlNet

URL or the path to the embedding weights.


            The tokens to map the embedding weights to. Use these tokens in your prompts.
        

Tokens

Embedding


            The value to set the image projection shortcut to. For FaceID plus V1 models,
            this should be set to False. For FaceID plus V2 models, this should be set to True.
            Default is True.
        

Image Projection Shortcut

URL or the path to the InsightFace model weights.

Insight Face Model Path

URL of the image to be used as the IP adapter.

Ip Adapter Image Url


            The mask to use for the IP adapter. When using a mask, the ip-adapter image size and the mask size must be the same
        

Ip Adapter Mask Url

Subfolder in the model directory where the IP adapter weights are stored.

Model Subfolder

URL or the path to the IP adapter weights.


            The scale of the IP adapter weight. This is used to scale the IP adapter weight
            before merging it with the base model.
        

Scale

Scale Json

The factor to apply to the unconditional noising of the IP adapter.

Unconditional Noising Factor

Weight Name

IPAdapter

ImageSize


            Skips part of the image generation process, leading to slightly different results.
            This means the image renders faster, too.
        

Clip Skip


            If set to true, the controlnet will be applied to only the conditional predictions.
        

Controlnet Guess Mode


            The control nets to use for the image generation. You can use any number of control nets
            and they will be applied to the image at the specified timesteps.
        

Controlnets

If set to true, the latents will be saved for debugging.

Debug Latents

If set to true, the latents will be saved for debugging per pass.

Debug Per Pass Latents


            The embeddings to use for the image generation. Only a single embedding is supported at the moment.
            The embeddings will be used to map the tokens in the prompt to the embedding weights.
        

Embeddings

If set to true, the safety checker will be enabled.

Enable Safety Checker

The eta value to be used for the image generation.


            The CFG (Classifier Free Guidance) scale is a measure of how close you want
            the model to stick to your prompt when looking for a related image to show you.
        

Guidance scale (CFG)


            The URL of the IC Light model image to use for the image generation.
        

Ic Light Image Url


            The URL of the IC Light model background image to use for the image generation.
            Make sure to use a background compatible with the model.
        

Ic Light Model Background Image Url


            The URL of the IC Light model to use for the image generation.
        

Ic Light Model Url


            The path to the image encoder model to use for the image generation.
        

Image Encoder Path


            The subfolder of the image encoder model to use for the image generation.
        

Image Encoder Subfolder


            The weight name of the image encoder model to use for the image generation.
        

Image Encoder Weight Name

Image Format

URL of image to use for image to image/inpainting.


            The IP adapter to use for the image generation.
        

Ip Adapter


            The LoRAs to use for the image generation. You can use any number of LoRAs
            and they will be merged together to generate the final image.
        

Loras

URL or HuggingFace ID of the base model to generate the image.

Model Name


            The negative prompt to use.Use it to address details that you don't want
            in the image. This could be colors, objects, scenery and even the small details
            (e.g. moustache, blurry, low resolution).
        

Negative Prompt

The amount of noise to add to noise image for image. Only used if the image_url is provided. 1.0 is complete noise and 0 is no noise.

Noise Strength


            Number of images to generate in one request. Note that the higher the batch size,
            the longer it will take to generate the images.
        

Number of images


            Increasing the amount of steps tells Stable Diffusion that it should take more steps
            to generate your final result which can increase the amount of detail in your image.
        

Search

Input

Result

Logs