如何使用labelme中的AI多边形（AI-polygon）标注

文章目录

• 1.创建labelme虚拟环境
• 2.下载AI标注模型
• 3.修改配置文件
• 4.愉快地使用labelme的AI标注工具

  1.创建labelme虚拟环境

  （1）创建基础环境并激活

  conda create -n labelme python=3.8
  conda activate labelme
  

  （2）安装labelme

  pip install labelme -i https://pypi.tuna.tsinghua.edu.cn/simple/ numpy
  

  （3）使用labelme启动

  

  如果是第一次装labelme，打开图像路径，右键图像后选择Create AI-Polygon，软件会自动下载并安装AI标注模型，我的下载速度太慢，导致第一次下载失败，最后选择了手动安装。

  

  2.下载AI标注模型

  可以选择在官网上下载AI自动标注模型下载地址

  

  如果连不到外网，可以通过迅雷网盘或者百度网盘提取模型

  迅雷网盘链接：https://pan.xunlei.com/s/VNkyiDkG9ORZRr7Mhx4ru3I8A1#

  提取码：2dbf

  百度网盘链接：https://pan.baidu.com/s/11xrWH4p_auHl-cKYjZ899Q?pwd=lg1j

  提取码：lg1j

  在anaconda虚拟环境中找到E:\programFiles\anaconda3\envs\labelme\Lib\site-packages\labelme此路径，将下载好的文件放入此文件夹下。

  

  3.修改配置文件

  （1）找到"E:\programFiles\anaconda3\envs\labelme\Lib\site-packages\labelme\ai\__init__.py"文件，并修改里面的模型路径。

  # flake8: noqa
  import logging
  import sys
  from qtpy import QT_VERSION
  __appname__ = "labelme"
  # Semantic Versioning 2.0.0: https://semver.org/
  # 1. MAJOR version when you make incompatible API changes;
  # 2. MINOR version when you add functionality in a backwards-compatible manner;
  # 3. PATCH version when you make backwards-compatible bug fixes.
  # e.g., 1.0.0a0, 1.0.0a1, 1.0.0b0, 1.0.0rc0, 1.0.0, 1.0.0.post0
  __version__ = "5.4.0a0"
  QT4 = QT_VERSION[0] == "4"
  QT5 = QT_VERSION[0] == "5"
  del QT_VERSION
  PY2 = sys.version[0] == "2"
  PY3 = sys.version[0] == "3"
  del sys
  from labelme.label_file import LabelFile
  from labelme import testing
  from labelme import utils
  import collections
  from .models.segment_anything import SegmentAnythingModel  # NOQA
  Model = collections.namedtuple(
      "Model", ["name", "encoder_weight", "decoder_weight"]
  )
  Weight = collections.namedtuple("Weight", ["url", "md5"])
  # MODELS = [
  #     Model(
  #         name="Segment-Anything (speed)",
  #         encoder_weight=Weight(
  #             url="https://github.com/wkentaro/labelme/releases/download/sam-20230416/sam_vit_b_01ec64.quantized.encoder.onnx",  # NOQA
  #             md5="80fd8d0ab6c6ae8cb7b3bd5f368a752c",
  #         ),
  #         decoder_weight=Weight(
  #             url="https://github.com/wkentaro/labelme/releases/download/sam-20230416/sam_vit_b_01ec64.quantized.decoder.onnx",  # NOQA
  #             md5="4253558be238c15fc265a7a876aaec82",
  #         ),
  #     ),
  #     Model(
  #         name="Segment-Anything (balanced)",
  #         encoder_weight=Weight(
  #             url="https://github.com/wkentaro/labelme/releases/download/sam-20230416/sam_vit_l_0b3195.quantized.encoder.onnx",  # NOQA
  #             md5="080004dc9992724d360a49399d1ee24b",
  #         ),
  #         decoder_weight=Weight(
  #             url="https://github.com/wkentaro/labelme/releases/download/sam-20230416/sam_vit_l_0b3195.quantized.decoder.onnx",  # NOQA
  #             md5="851b7faac91e8e23940ee1294231d5c7",
  #         ),
  #     ),
  #     Model(
  #         name="Segment-Anything (accuracy)",
  #         encoder_weight=Weight(
  #             url="https://github.com/wkentaro/labelme/releases/download/sam-20230416/sam_vit_h_4b8939.quantized.encoder.onnx",  # NOQA
  #             md5="958b5710d25b198d765fb6b94798f49e",
  #         ),
  #         decoder_weight=Weight(
  #             url="https://github.com/wkentaro/labelme/releases/download/sam-20230416/sam_vit_h_4b8939.quantized.decoder.onnx",  # NOQA
  #             md5="a997a408347aa081b17a3ffff9f42a80",
  #         ),
  #     ),
  # ]
  MODELS = [
      Model(
          name="Segment-Anything (speed)",
          encoder_weight=Weight(
              url="E:\programFiles\\anaconda3\envs\labelme\Lib\site-packages\labelme\model_file\sam_vit_b_01ec64.quantized.encoder.onnx",  # NOQA
              md5="80fd8d0ab6c6ae8cb7b3bd5f368a752c",
          ),
          decoder_weight=Weight(
              url="E:\programFiles\\anaconda3\envs\labelme\Lib\site-packages\labelme\model_file\sam_vit_b_01ec64.quantized.decoder.onnx",  # NOQA
              md5="4253558be238c15fc265a7a876aaec82",
          ),
      ),
      Model(
          name="Segment-Anything (balanced)",
          encoder_weight=Weight(
              url="E:\\programFiles\\anaconda3\\envs\\labelme\\Lib\\site-packages\\labelme\\model_file\\sam_vit_l_0b3195.quantized.encoder.onnx",  # NOQA
              md5="080004dc9992724d360a49399d1ee24b",
          ),
          decoder_weight=Weight(
              url="E:\\programFiles\\anaconda3\\envs\\labelme\\Lib\\site-packages\\labelme\\model_file\\sam_vit_l_0b3195.quantized.decoder.onnx",  # NOQA
              md5="851b7faac91e8e23940ee1294231d5c7",
          ),
      ),
      Model(
          name="Segment-Anything (accuracy)",
          encoder_weight=Weight(
              url="E:\\programFiles\\anaconda3\\envs\\labelme\\Lib\\site-packages\\labelme\\model_file\\sam_vit_h_4b8939.quantized.decoder.onnx",  # NOQA
              md5="958b5710d25b198d765fb6b94798f49e",
          ),
          decoder_weight=Weight(
              url="E:\\programFiles\\anaconda3\\envs\\labelme\\Lib\\site-packages\\labelme\\model_file\\sam_vit_h_4b8939.quantized.encoder.onnx",  # NOQA
              md5="a997a408347aa081b17a3ffff9f42a80",
          ),
      ),
  ]
  

  （2）找到E:\programFiles\anaconda3\envs\labelme\Lib\site-packages\labelme\widgets\canvas.py文件夹并修改initializeAiModel方法

      def initializeAiModel(self, name):
          if name not in [model.name for model in labelme.ai.MODELS]:
              raise ValueError("Unsupported ai model: %s" % name)
          model = [model for model in labelme.ai.MODELS if model.name == name][0]
          if self._ai_model is not None and self._ai_model.name == model.name:
              logger.debug("AI model is already initialized: %r" % model.name)
          else:
              logger.debug("Initializing AI model: %r" % model.name)
              self._ai_model = labelme.ai.SegmentAnythingModel(
                  name=model.name,
                  # encoder_path=gdown.cached_download(
                  #     url=model.encoder_weight.url,
                  #     md5=model.encoder_weight.md5,
                  # ),
                  # decoder_path=gdown.cached_download(
                  #     url=model.decoder_weight.url,
                  #     md5=model.decoder_weight.md5,
                  # ),
                  encoder_path=model.encoder_weight.url,
                  decoder_path=model.decoder_weight.url,
              )
          self._ai_model.set_image(
              image=labelme.utils.img_qt_to_arr(self.pixmap.toImage())
          )
  

  4.愉快地使用labelme的AI标注工具

  这样再激活虚拟环境，使用labelme命令打开标注工具，右键选择AI标注，双击标注完成。

  

  参考链接：labelme加载AI模型