ComfyUI Segment Anything

Overview

The ComfyUI Segment Anything repository is a collection of custom nodes designed for the ComfyUI platform. This project implements core functionalities from the sd-webui-segment-anything repository, ensuring consistency in output for identical inputs. The primary objective of this repository is to integrate advanced segmentation features within the ComfyUI framework.

Repository Purpose

This repository facilitates image segmentation tasks in ComfyUI by providing essential models and nodes that enable users to perform segmentation operations. The integration of these nodes allows for advanced manipulations, including automatic downloading and configuration of segmentation models, making it a comprehensive tool for users who require intricate image processing capabilities.

Installation Instructions

To install and set up the ComfyUI Segment Anything repository, follow these steps:

1. Clone the Repository

   • Begin by cloning the repository to your local machine.

2. Install Python Dependencies

   • Ensure that all necessary Python dependencies are installed by running the following command:

     pip3 install -r requirements.txt
     

3. Model Setup

   • The essential models for this repository will be automatically downloaded as needed. If you encounter slow download speeds, configure your environment to use a proxy by setting the HTTP_PROXY and HTTPS_PROXY environment variables.
   • Alternatively, you can manually download the models using the provided links in the repository and place them in the appropriate directories within the ComfyUI root directory.

Node Descriptions

The repository provides several custom nodes, each designed to perform specific segmentation tasks:

1. IsMaskEmpty

   • Implemented as IsMaskEmptyNode in __init__.py.
   • Function: Determines if a segmentation mask is empty.

2. SAMModelLoader

   • Implemented as SAMModelLoader in __init__.py.
   • Function: Loads models for segmentation tasks using the "Segment Anything" methodology.

3. GroundingDinoModelLoader

   • Implemented as GroundingDinoModelLoader in __init__.py.
   • Function: Loads GroundingDino models for segmentation purposes.

4. GroundingDinoSAMSegment

   • Implemented as GroundingDinoSAMSegment in __init__.py.
   • Function: Performs segmentation combining GroundingDino models with the "Segment Anything" methodology.

5. InvertMask

   • Implemented as InvertMask in __init__.py.
   • Function: Inverts a given segmentation mask.

Special Features

• Automatic Model Downloading: The repository is designed to automatically manage the downloading and configuration of necessary models, easing the setup process.
• Proxy Support: Users experiencing slow download speeds can configure proxy settings to enhance download efficiency.

Utility in ComfyUI Workflows

The ComfyUI Segment Anything repository enriches the ComfyUI environment by introducing advanced segmentation capabilities. By offering a suite of nodes tailored for various segmentation tasks, this repository empowers users to integrate image segmentation seamlessly into their workflows. This integration allows users to process and manipulate images with precision, making it an invaluable resource for tasks requiring image segmentation within the ComfyUI framework.

Contribution

Contributions to the ComfyUI Segment Anything project are highly welcomed. Developers interested in enhancing the repository are encouraged to fork the project, make improvements, and submit pull requests for review and potential inclusion in the main codebase. Every contribution, no matter how minor, is appreciated and contributes to the evolution of the project.