Input

Supported Image Formats

WormLib supports three microscopy file formats:

DeltaVision (.dv), Nikon ND2 (.nd2), and TIFF (.tif, .tiff) - Multi-channel volumetric images (Z-stacks) - Native support for 4D data (X, Y, Z, time) and channel extraction

## DeltaVision files follow a specific naming pattern for reference vs. color images:

Critical Pattern:

_R3D_REF  → brightfield reference (2D)
_R3D      → color/fluorescence (4D multi-channel)

Reference image (brightfield):

Contains: - 2D brightfield transmission image for segmentation - Single channel per Z-slice

Color image (fluorescence):

Contains: - 4D (C,Z,Y,X) - 4 channels (Cy5, mCherry, FITC, DAPI) × Z-slices - Each channel is a separate index (0, 1, 2, 3)

Do NOT confuse these! The _R3D file must not include _D3D in the name and must be a different file from _REF.

—

File Organization Best Practice

Give your files informative names at the time of acquisition. Example: 230713_Lp306_L4440_11_R3D_REF.dv = date_strain_condition_replicate/. Organize your data in a clear folder structure:

data>image_subdirectory>image_files

data/
├── 230713_Lp306_L4440_11/
│   ├── 230713_Lp306_L4440_11_R3D_REF.dv
│   └── 230713_Lp306_L4440_11_R3D.dv
├── 230713_Lp306_L4440_12/
│   ├── 230713_Lp306_L4440_12_R3D_REF.dv
│   └── 230713_Lp306_L4440_12_R3D.dv
└── 230713_Lp306_L4440_13/
    ├── 230713_Lp306_L4440_13_R3D_REF.dv
    └── 230713_Lp306_L4440_13_R3D.dv

Benefits: - Both reference and color files in same folder allows WormLib to auto-detect and load both - Easy batch processing - Image output subdirectory automatically created in the same folder as input images —

Loading Images in Code

Automatic file type detection. DeltaVision, Nikon and TIFF files supported :

import wormlib

# Path to image subdirectory
image_path = Path("data/230713_Lp306_L4440_11")

result = wormlib.load_images(
    image_path=str(image_path),
    output_directory="output/",
    channel_names={
        'Cy5': 'set3_mRNA', # Describe what mRNA is in this channel
        'mCherry': 'erm1_mRNA', # Describe what mRNA is in this channel
        'FITC': 'membrane', # Describe what marker is in this channel
        'DAPI': 'DAPI',
        'brightfield': 'brightfield'
    },
    channel_indices={
        'Cy5': 0,
        'mCherry': 1,
        'FITC': 2,
        'DAPI': 3,
        'brightfield': None
    }
)

WormLib automatically detects .dv extension and finds both _R3D_REF and _R3D files in the same folder. It can load brightfield from the reference file (R3D_REF.dv) and channels from the color file (R3D.dv) and returns organized image data.

Result structure:

Image Data Dictionary
Key	Value/Type	Description
image_type	str	Image format type (e.g., ‘DeltaVision’)
image_name	str	Image filename without extension (e.g., ‘230713_Lp306_L4440_11’)
bf	numpy array	Brightfield (transmission) 2D image (1024, 1024)
image_Cy5	numpy array	Channel 0 max projection (Cy5 fluorophore)
image_mCherry	numpy array	Channel 1 max projection (mCherry fluorophore)
image_FITC	numpy array	Channel 2 max projection (FITC/GFP fluorophore)
image_nuclei	numpy array	Channel 3 max projection (DAPI/nuclei stain)
Cy5_array	numpy array	Channel 0 full 3D volume (Z, Y, X)
mCherry_array	numpy array	Channel 1 full 3D volume (Z, Y, X)
FITC_array	numpy array	Channel 2 full 3D volume (Z, Y, X)
nuclei_array	numpy array	Channel 3 full 3D volume (Z, Y, X)
grid_width	int	Grid width in pixels (80)
grid_height	int	Grid height in pixels (80)

—