This project provides a comprehensive toolset for processing, analyzing, and aligning tissue images from OME-TIFF files. It includes features for image preprocessing, tissue detection, cropping, and alignment.
Clone this repository and navigate into the project directory. Ensure you have Python 3.8 or later installed. It's recommended to use a virtual environment.
git clone [email protected]:hubmapconsortium/SectionAligner.git
cd SectionAligner
pip install -r requirements.txtThe main script can be run from the command line, providing various options for processing the images:
python main.py --input_path "path/to/your/image.ome.tiff" --output_dir "path/to/output" --num_tissue 8 --pixel_size [0.5073519424785282, 0.5073519424785282] ---crop_only False--num_tissue: Number of tissues to detect (default is 8).
--pixel_size: Pixel size in microns. (IMPORTANT as it determines how much to downsample and will effect end results)
--crop_only: Only identify tissues and crop, without alignment (default is False).
--num_tissue: The number of tissues to detect. Default is 8.
--level: Pyramid level of the image. Default is 0, which is the original image size.
--thresh: Threshold value for binarization. If not set, Otsu's method is used.
--kernel_size: Size of the structuring element used for morphological operations. Default is 100.
--holes_thresh: Area threshold for removing small holes. Default is 5000.
--scale_factor: Scale factor for downscaling images. Default is 8.
--padding: Padding for bounding boxes during cropping. Default is 50.
--connect: Connectivity for connected components. Default is 2.
--pixel_size: Physical pixel size of the image in microns. Default is [0.5073519424785282, 0.5073519424785282].
--output_dir: Output directory for saving images. Default is ./outputs.
--input_path: Input path for reading images. Default is the specified path to an OME-TIFF file.
--output_file_basename: Basename for output files. Default is aligned_tissue.
--align_upsample_factor: Upsample factor for aligning images. Default is 2.
--optimize: Whether to optimize alignment parameters using Optuna. Boolean. Default is True.
--crop_only: If True, only identify tissues and crop, without aligning. Boolean. Default is False.
Tissue Detection: Identifies tissues within OME-TIFF images.
Image Cropping: Crops identified tissues for focused analysis.
Image Alignment: Aligns tissue slices for better comparison and analysis.
Optimization: Uses Optuna for optimizing image alignment parameters.
Ted Zhang [[email protected]]
Bob Murphy [[email protected]]