如何创建TIFF使QuPath点击放大按钮时提示输入放大倍数？

我用Python的tifffile库创建TIFF图像，元数据设置及代码如下。但在QuPath中打开生成的文件后，点击“0.01x”（放大倍数）按钮时，系统提示输入“downsampling factor（下采样因子）”而非“magnification（放大倍数）”。我知道这可能是图像未校准导致，但已经在元数据中提供了像素尺寸，请问要修改代码何处才能让系统提示输入放大倍数？

原代码 #

panorama_metadata = {
"MetaMorph": "no",
"BitsPerSample": "8",
"Orientation": "1st row - top; 1st column - left",
"TileWidth": self.tile_width,
"Compression": self.compression_method,
"PhotometricInterpretation": "YCbCr",
"ResolutionUnit": "micrometer",
"XResolution" : 1/285,
"YResolution" : 1/285,
"ReferenceBlackWhite": "0",
"NumberOfChannels": "3",
"SamplesPerPixel": "3",
"ImageWidth": total_pixels_in_cols,
"TileLength": self.tile_height,
"MetaDataPhotometricInterpretation": "RGB",
"SampleFormat": "unsigned integer",
"ImageLength": total_pixels_in_rows,
"PlanarConfiguration": "Chunky",
}

with TiffWriter(os.path.join(self.output_path,
self.slide_name + ".ome.tif"),
bigtiff=True) as tif:
                tif.save(data=self.get_numpy_arrays(number_of_columns,
                                                    number_of_rows,
                                                    max_folder_path),
                         shape=panorama_shape,
                         tile=(self.tile_height, self.tile_width),
                         dtype=numpy.uint8, photometric='rgb',
                         metadata=metadata)

关键修改点 #

1. 修正元数据变量名错误
   你定义的元数据字典是panorama_metadata，但保存时传入的是metadata，这直接导致设置的元数据完全没生效。把metadata=metadata改成metadata=panorama_metadata。

2. 添加物镜放大倍数元数据
   QuPath需要明确的物镜放大倍数信息才能识别“放大倍数”选项，在panorama_metadata中添加：

   "ObjectiveLensPower": 20,  # 替换为你的实际物镜倍数，比如10、20、40等
   

3. 确保OME-TIFF格式规范
   你保存的是.ome.tif，但需要启用tifffile的OME格式支持，在tif.save()中添加ome=True，让QuPath更好地识别OME标准的元数据。

4. 补充物理尺寸校准信息（可选）
   虽然你设置了分辨率，但补充更直观的物理尺寸字段能提升QuPath的识别准确性，在panorama_metadata中添加：

   "PhysicalSizeX": 0.285,  # 对应285纳米，即1/285微米
   "PhysicalSizeXUnit": "µm",
   "PhysicalSizeY": 0.285,
   "PhysicalSizeYUnit": "µm"
   

修正后的代码示例 #

import numpy as np
from tifffile import TiffWriter

# 假设self相关变量已定义
panorama_metadata = {
    "MetaMorph": "no",
    "BitsPerSample": "8",
    "Orientation": "1st row - top; 1st column - left",
    "TileWidth": self.tile_width,
    "Compression": self.compression_method,
    "PhotometricInterpretation": "YCbCr",
    "ResolutionUnit": "micrometer",
    "XResolution": 1/285,
    "YResolution": 1/285,
    "ReferenceBlackWhite": "0",
    "NumberOfChannels": "3",
    "SamplesPerPixel": "3",
    "ImageWidth": total_pixels_in_cols,
    "TileLength": self.tile_height,
    "MetaDataPhotometricInterpretation": "RGB",
    "SampleFormat": "unsigned integer",
    "ImageLength": total_pixels_in_rows,
    "PlanarConfiguration": "Chunky",
    # 添加物镜放大倍数
    "ObjectiveLensPower": 20,
    # 补充物理尺寸信息
    "PhysicalSizeX": 0.285,
    "PhysicalSizeXUnit": "µm",
    "PhysicalSizeY": 0.285,
    "PhysicalSizeYUnit": "µm"
}

with TiffWriter(os.path.join(self.output_path, self.slide_name + ".ome.tif"), bigtiff=True) as tif:
    tif.save(
        data=self.get_numpy_arrays(number_of_columns, number_of_rows, max_folder_path),
        shape=panorama_shape,
        tile=(self.tile_height, self.tile_width),
        dtype=np.uint8,
        photometric='rgb',
        metadata=panorama_metadata,  # 修正变量名
        ome=True  # 启用OME-TIFF格式
    )

内容的提问来源于stack exchange，提问作者Monalisha Bhowmik