Comfyui-Yolov8-JSON

Comfyui-Yolov8-JSON
★ 28

目标检测实例分割JSON输出自动加载模型
基于Yolov8的ComfyUI节点,进行目标检测与分割,自动加载模型并输出图片、掩码及Labelme JSON。
💡 在ComfyUI中批量检测并导出掩码与Labelme JSON
🍴 10 Forks💻 Python🔄 2024-08-28
📦
网盘下载
复制链接后前往夸克网盘下载
https://pan.quark.cn/s/9671236b7e59
📦 requirements.txt
ultralytics
numpy
torch
PIL
📄 README

Comfyui-Yolov8-JSON

This node is mainly based on the Yolov8 model for object detection, and it outputs related images, masks, and JSON information.

README

  • English
  • 简体中文
  • NODES

    |name |description |

    |—————————–|—————————————————————-|

    |Load Yolov8 Model |Default, select the Yolov8 model, supports automatic download |

    |Load Yolov8 Model From Path |Load the model from the specified path |

    |Apply Yolov8 Model |Apply Yolov8 detection model |

    |Apply Yolov8 Model Seg |Apply Yolov8 segmentation model |

    |Save Labelme Json |Save the original image and the corresponding Labelme JSON format file to the output directory, with support for custom directories |

    |Draw Labelme Json |Based on the Labelme JSON, draw recognition boxes on the image and output masks, with support for specifying the labels to display, changing colors, and renaming labels |

    INSTALL

    If you need to display JSON formatted data or save it, you need to install the Comfyui-Toolbox node in advance.

    This node calls the official Python package, and you also need to install the following dependencies:

    pip install -r requirements.txt

    MODEL

    This node supports automatic model downloads.

    You can manually download the models to the specified folder models/yolov8 from Yolov8 as shown below:

    ComfyUI
        models
            yolov8
                yolov8l.pt
                yolov8s.pt
                yolov8n.pt
                yolov8m.pt
                yolov8x.pt
                yolov8l-seg.pt
                yolov8s-seg.pt
                yolov8n-seg.pt
                yolov8m-seg.pt
                yolov8x-seg.pt