Image-related record format
Image processing record (IPR)
Image processing record (IPR) file is designed to record the whole-life information of a microscope image from photo-taking to processing. Each staining image (ssDNA/DAPI, IF, H&E) includes six basic groups, "ImageInfo", "QCInfo", "Stitch", "TissueSeg", "CellSeg", and "Register", which are used to store microscopy photo-taking information, image quality control information, image stitching records, tissue segmentation records, cell segmentation records (optional), and registration records.
Attribute
Description
IPRVersion
IPR file format version.
Dateset
Description
Preview
A 2D matrix merges stitched image, tissue segmentation boundary and cell segmentation boundary.
/<stainType>/ImageInfo: Group records basic image information.
Attributes
Description
AppFileVer
Microscope software version.
BackgroundBalance
Background balance.
BitDepth
The bit-depth of a camera sensor describes its ability to transform the analog signal coming from the pixel array into a digital signal.
Brightness
Relative intensity affecting a person or sensor.
ChannelCount
Number of RGB channels.
ColorEnhancement
Whether enhanced image color display or not.
Contrast
The difference in color and intensity of the depicted object from its background.
DeviceSN
Microscope device serial number.
DistortionCorrection
Whether fixed distortion or not.
ExposureTime
Exposure time in ms.
FOVHeight
Height of an individual FOV in pixel.
FOVWidth
Width of an individual FOV in pixel.
Gain
Amplification applied to the signal by the image sensor.
Gamma
The coefficient links between the human eye and the digital camera.
GammaShift
Whether adapt the digital image taken with the help of a linearly recording camera to the nonlinear perception of the human eye or not.
Illuminance
Intensity of light.
Manufacturer
Microscope manufacturer.
Model
Microscope model.
Overlap
Overlapping pixels between single tiles.
PixelSizeX
Size of pixel in x direction.
PixelSizeY
Size of pixel in y direction.
QCResultFile
Prefix of ImageQC/ImageStudio result file, the unique identifier of the image.
ScanChannel
Fluorescence channel.
ScanCols
Number of columns scanned.
ScanObjective
Magnification power of the scan objective lens.
ScanRows
Number of rows scanned.
ScanTime
Scan date and time.
Sharpness
Degree of clarity of the edge(s) of the image.
RegisterVersion
Image auto-processing & matrix registration software version.
StitchedImage
Whether the corresponding image is a panorama image (true) or a set of tiled images (false).
STOmicsChipSN
Stereo-seq Chip serial number.
WhiteBalance
An adjustment in electronic and film imaging that corrects the color balance of the lighting.
DatasetDataType: 1D array
Description
RGBScale
RGB color.
/<stainType>/QCInfo: Group records the QC information of the image.
Attributes
Description
ClarityScore
Reference score for evaluating the clearness of cell boundaries.
Experimenter
Email of the experimenter who did QC for the image.
GoodFOVCount
Number of FOVs that have identified more than 3 track cross points.
ImageQCVersion
Image QC version.
QCPassFlag
Whether the corresponding image passed QC.
RemarkInfo
Any remarks, notes, comments on the image.
StainType
Stain type of microscopy image.
TotalFOVCount
Total number of FOVs.
TrackLineScore
Reference score for evaluating whether the detected track lines can be used as references for image stitching and registering with gene expression matrix. (This score only evaluate whether the program detected track lines on the image, it does not infer the clarity of the lines or the images).
TrackLineChannel
Shooting channel of track lines.
TrackCrossQCPassFlag
Whether the QC of the track-cross point is passed.
ScopeStitchQCScore
Score to assess the stability of microscope overlap (DAPI & mIF only).
ScopeStitchQCPassFlag
Whether the QC of microscope stitch is passed (DAPI & mIF only).
TemplateValidArea
Proportion of encircled area by the detected points that match the global track line template (within the error range of 5 pixels), to the entire image area.
TemplateRecall
Proportion of detected points, which match the global track line template (within the error range of 5 pixels), to the points derived from the chip track line rules.
ClarityCutSize
Size of cut images (based on FOV) for clarity evaluation.
ClarityCounts
Counts of cut images in different categories.
ClarityOverlap
Overlap of cut images for clarity evaluation.
DatasetDataType: 2D array
Description
ScopeStitchQCMatrix
Matrix of overlap deviation of each microscope FOV (DAPI & mIF only).
CrossPoints/row_col*n
Group of datasets for each FOV that records the track cross point coordinates. (Row and col stand for the FOV row and column index number, and n stands for number of FOVs). Each dataset is a 2D array record, (x, y) coordinates of track cross points in each FOV.
/<stainType>/Calibration: Group records the calibration information (<protein>_IF only).
Attributes
Description
CalibrationQCPassFlag
Whether the calibration QC of IF image is passed.
/<stainType>/Calibration/Scope:
Attributes
Description
Confidence
Calibration confidence of microscope-stitched tiled image vs. IF.
OffsetX
Horizontal offset.
OffsetY
vertical offset.
[optional]/<stainType>/Calibration/Stereo:
Attributes
Description
Confidence
Calibration confidence of Stereo-stitched image vs. IF.
OffsetX
Horizontal offset.
OffsetY
Vertical offset.
/<stainType>/Stitch: Group records the stitching information.
Attributes
Description
StitchingScore
Reference score for stitching.
TemplateSource
The reference FOV for deriving the template used for rotating and scaling the microscopic images.
WhichStitch
Stitching method for image, including microscope, template and ripple stitching. The default for stitched image is microscope.
DatasetDataType: 2D array
Description
TemplatePoint
Center coordinates for deriving template lines.
TransformTemplate
Coordinates of template points registered with expression matrix.
/<stainType>/Stitch/StereoStitch: Group records the image stitching information processed by Stereo program.
Attributes
Description
StitchedGlobalHeight
Height of stitched tiled images using Stereo stitching algorithm. Tiled image only.
StitchedGlobalWidth
Width of stitched tiled images using Stereo stitching algorithm. Tiled image only.
DatasetDataType: 2D array
Description
StitchedGlobalLoc
Coordinates for the Stereo stitched tiled image. Tiled image only.
/<stainType>/Stitch/ScopeStitch: Group records the image stitching information processed by microscope imaging software.
Attributes
Description
GlobalHeight
Height of panorama image.
GlobalWidth
Width of panorama image.
DatasetDataType: 2D array
Description
GlobalLoc
Coordinates for the stitched tiled image (either program stitched or microscope stitched).
ScopeJitterDiff
Jitter offset of microscope stitching.
ScopeHorizontalJitter
Horizontal jitter offset of microscope stitching. Tiled image only.
ScopeVerticalJitter
Vertical jitter offset of microscope stitching. Tiled image only.
/<stainType>/Stitch/StitchEval: Group records the evaluation result of stitching.
Attribute
Description
MaxDeviation
Maximum stitching deviation.
DatasetDataType: 2D array
Description
GlobalDeviation
Global stitching deviation matrix. Tiled image only.
StitchEvalH
Stitching deviation matrix for the horizontal axes.
StitchEvalV
Stitching deviation matrix for the vertical axes.
/<stainType>/Register: Group records the information that aligns images with gene expression matrix.
Attributes
Description
CounterRot90
Count of counter-clockwise rotation of 90 degrees.
Flip
Whether horizontally flipped or not.
MatrixShape
Height and width of the gene expression matrix.
OffsetX
Offset between microscope image and gene expression matrix in x-axis.
OffsetY
Offset between microscope image and gene expression matrix in y-axis.
RegisterScore
Reference score for registration.
Rotation
Rotation degree between raw image and deviation template.
ScaleX
Scale between raw image and deviation template in horizontal direction.
ScaleY
Scale between raw image and deviation template in vertical direction.
XStart
Gene expression matrix offset x (GEF geneExp/binN/expression attribute minX).
YStart
Gene expression matrix offset y (GEF geneExp/binN/expression attribute minY).
ManualRotation
Manual rotation degree of the raw image around the center point.
ManualScaleX
Manual scale of the raw image in horizontal direction based on center point.
ManualScaleY
Manual scale of the raw image in vertical direction based on center point.
DatasetDataType: 2D array
Description
MatrixTemplate
List of track cross points derived from gene expression matrix.
/<stainType>/TissueSeg: Group records the tissue segmentation information.
Attributes
Description
TissueSegScore
Reference score for tissue segmentation.
TissueSegShape
Image shape for tissue segmentation mask image.
Threshold
The threshold for tissue segmentation in IF images is divided into maximum and minimum values (IF only).
DatasetDataType: 2D array
Description
TissueMask
Encoded tissue segmentation mask file (after registration with gene expression matrix).
/<stainType>/CellSeg: Group records the cell segmentation information.
Attributes
Description
CellSegShape
Image shape for cell segmentation mask image.
DatasetDataType: 2D array
Description
CellMask
Encoded cell segmentation mask file (after registration with gene expression matrix).
CellSegTrace
Cell contour attributes, including height, width and area.
/StereoResepSwitch: Group stores the state of each module that whether the module needs to be performed.
Attributes
Description
stitch
Switch for performing stitching.
tissueseg
Switch for performing tissue segmentation.
cellseg
Switch for performing cell segmentation.
register
Switch for performing registration.
/ManualState: Group stores the state of each module that whether the module has been manually processed.
Attributes
Description
stitch
Whether manually stitched the tiled images.
tissueseg
Whether manually delineated the tissue coverage region.
cellseg
Whether manually delineated the cell coverage regions.
register
Whether manually aligned microscope image and gene expression matrix.
calibration
Whether manually calibrated two images. For example, manually match IF image with DAPI image.
Recorded image processing (RPI)
The image pyramid model is a multi-resolution hierarchical model that is used to store and display images in different resolutions. For the same field of view, the layer of the image pyramid that is closest to the bottom includes the most detailed information and has the largest scale. register pipeline performs the down-sampling step on the registered image, and the result images are layered to construct a pyramid with the suffix ".rpi". For each resolution layer, the intact registered image is split into 256 pixels x 256 pixels tiles. If the size of a layer is smaller than 256 x 256, the image will then remain intact. The outer layer group of RPI file is defined according to its stain type, generally including ssDNA, DAPI, IF (immunofluorescence image especially for protein, group name as <protein_IF>), and H&E. In each group of stain types, multiple image results, including an image subgroup (registered microscopy image), a TissueMask subgroup (registered mask boundary for the tissue coverage area), and a CellMask subgroup (registered mask boundaries for the cell coverage area, optional), could be saved respectively.
Schematic diagram of the image pyramid:
Attributes
Description
imgSize
Size of the image.
sizex
Max of x.
sizey
Max of y.
version
RPI file version.
x_start
Offset x of the matrix.
y_start
Offset y of the matrix.
/<stainType>/<imageLayer>: Group records the specific layer information, mainly including "Image", "TissueMask", "CellMask" and "CellMask_adjusted".
Attributes
Description
MaxGrayLevel
Maximum gray value.
MinGrayLevel
Minimum gray value.
GrayLevelElbow
Elbow point of gray value.
TrackLayer
Whether track layer.
Color
Layer color.
Groups
Description
binN
Detailled downsampling metadata for each bin size.
Last updated
