AddClassAttributeFromARFFimageAndMasks

Reads the image ARFF file, and all provided mask ARFF files. Outputs N rows for each class. Requires less memory, as all input ARFF files are read row by row, so only one row of each input ARFF needs to stay in memory. In addition sets of valid row indices for each mask file need to be constructed, to allow random selection of random samples. From the mask ARFF files, the correct class of the current image row is detected, and the according class information is added at the end of each image ARFF line. The input mask files need to be read two times, the first time the number of samples in each class are determined, so that in the main run the correct number of samples may be selected randomly from the class rows. At least two mask arff files should be provided (though one would be enough, but then all image rows would belong to a single class.

@param class sample size N
@param output arff file
@param image arff file
@param mask arff file 0
@param mask arff file 1
@param mask arff file 2
@param [..]
@return ARFF file

ApplyClass0ToGrayScaleImage

Converts classified ARFF result file to an image by using a labeled mask image.

@param input filename(s)
@return image file

ApplyClass0ToImage

Converts classified ARFF result file to an image by using a labeled mask image.

@param input filename(s) for ARFF files
@return visualization image of class 0 from input file

ApplyMask

Converts classified ARFF result file to an image by using a labeled mask image.

@param directory with input file(s) images
@param directory with input file(s) images ROIs
@return cleared image (by given ROI)

ApplyMaskImage

Applies a mask on the input image, the output will be saved in a separate target file.

@param image path to input image
@param mask path to mask image
@param target path to output directory
@return image file

ApplyROIAndCrop

Apply a mask to given an image and return a cropped version with specified name (target file) and extension (file format).

@param image
@param mask
@param target file name
@param file format
@return image file

ArffFromImageFileGenerator

Converts image into an ARFF file.

@param class-count
@param input filename
@return ARFF file

ArffSampleFileGenerator

Sample Extraction from input images, generates .arff file for classifier training.

@param class-count (negative in case of foreground/background segmentation)
@param sample-size
@param input filename(s)
@return ARFF file

ArffToImageFileGenerator

Converts classified ARFF result file to an image by using FG mask.

@param channel count (int)
@param input ARFF file(s)
@return classified image

ArffToProbabilityImageFileGenerator

Converts classified ARFF result file (including pixel-probabilities for each class) to an grayscale-image (use of FG mask is possible).

@param channel count (positive value = .png output, negative value = .tif output)
@param threshold for acceptance
@param input filename(s) of ARFF file(s)
@return grayscale image

ClacImage

Create image by calculating the difference, sum, division or multiplication of two images.

@param image A
@param image B
@param output file
@param one of this: +,-,*,/, absdiff, and
@param RGB or Float mode
@return difference image according to operation mode

CalculateDistanceMap

Calculate distance map float image. Modes: 0 - eucl. distance map X, 1 - eucl. distance map Y, 2 - eucl. distance map dist, 3 - euchl. distance map degree*255/360.

@param mode 0/1/2/3
@param input mask
@param target TIFF file (distance map)
@return distance map image (tif)

ColoredRegionGrowing

Nearest-neighbor approach, to find colorized areas near uncolored foreground area. Assign color of nearest colorized pixel to uncolored pixels. Input is a colored image, with some uncolored (white) areas, mask image, background pixels (black) are ignored during processing.

@param input image (RGB)
@param image for over-drawing (edges)
@param target filename
@return result image

ColoredRegionGrowingSingle

Nearest-neighbor approach, to find colorized areas near uncolored foreground area. Assign color of nearest colorized pixel to uncolored pixels. Input is a colored image, with some uncolored (white) areas, mask image, background pixels (black) are ignored during processing.

@param input image (RGB)
@param target filename
@return result image

CombineDirectionalProbabilityImages

Combines two probability images.

@param image 1 (grayscale tif)
@param image 2 (grayscale tif)
@param taarget filename
@param Modes: 0 - img1 = X, img2 = Y; 1 - img1 = DIST, img2 = DEGREE*255/360
@return image file

ConvertPSIIToTif

Command to convert .dumm (raw image files for PSII measurements) and .fimg (result image files, including the results for a feature as calculated during PSII analysis) to .tif image files.

@param path to folder including files for conversion
@return converted images (saved into same folder as used for input)

CreateCircularGradientImage

Creates an circular gradient image (tif float), the most far point from the center has value 0, the center point has value 1. The dimensions of the target image are determined from the given input (template) image.

@param input image
@param filename output (should have tif extension)
@return image file

CreateDiffImage

Create difference image, indicating true positive and true negative in white. False positive in blue and false negative in red.

@param ground truth mask
@param prediction
@param output file name for difference image
@return difference image

Erode

This command performs the erode operation on a given input image (black pixels will be regarded as foreground, white as background).

@param input image
@return image

ExportImagesFromHyperSpec

Creates separate images from hyper-spectral image data set (BSQ, BIL format).

@param prefix for output filename
@param overflow threshold or negative value to disable
@param filenames (input)
@return image files (channel images)

FILTER

Based on an input image an image operation is performed and a target result file is saved.

@param input image
@param target file
@param mask size (int)
@param sigma for gaussian blur (double)
@param operation mode (SHARPEN, BLUR, MEDIAN, TEXTURE, HARLICK, ALL)
@return filtered result image

MakeRGBComposite

This command combines three images 'channel_0, channel_1, channel_2' (0 = red, 1 = green, 2 = blue) to an RGB composite image.

@param input R/G/B images ('channel_0, channel_1, channel_2' (0 = red, 1 = green, 2 = blue))
@return RGB composite image

MergeArffFiles

Merges two or more ARFF files (concatinates the column data)

@param target file
@param input filenames
@param -ColIndex for ARFF files to be merged specified
@param optionally specific columns may be removed from the output (1...x)
@param by adding +str an string can be added into the last column! Return Code 1"
@return merged ARFF file

PoersetGenerator

Generates inverted mask of given masks (AND + INV). Also generates a powerset of the input combinations of images. All these images are then saved.

@param class count (int)
@param input image(s)
@return powerset combinations

QuantifyEnhanced

Compared to the simple Quantify command (works for one object in an image), it is possible to analyze and distinguish several separated objects in an image. The results are saved into a .csv file and a debug image.

@param output mode (0 = percentage, 1 = absolute values)
@param image file
@return csv file
@return debug image

Quantify

Reads and image and quantifies (counts) the foreground pixels, marked with different colors. For each color a the corresponding infection rate is calculated.

@param output mode (0 = percentage, 1 = absolute values)
@param image file (starting with classified_, cluster or ends with _cluster)
@return csv file

RGB2ALL

Create color channel files in diverse color spaces.

@param R/G/B input images (split channels)
@param color channel index (0..19, -1 = all, ? = list)
@param input color space index (0..4, -1 = all, ? = list)
@return channel_xyz_x.png, ... (diverse set of color channels)

RGB2Bayer

Create Bayer patterned image file (gray scale).

@param R/G/B image files
@return bayer_pattern image (tif)

RGB2HSB

Create H/S/B image files from R/G/B image files.

@param R/G/B image files
@return channel_hsv_h.png, channel_hsv_s.png, channel_hsv_v.png

RGB2LAB

Create L/a/b image files from R/G/B files.

@param R/G/B image files
@return channel_lab_l.png, channel_lab_a.png, channel_lab_b.png

RGB2XYZ

Transform R/G/B input image files to the corresponding X/Y/Z image files.

@param R/G/B image files
@return channel_xyz_x.png, channel_xyz_y.png, channel_xyz_z.png

RgbBayer2Arff

Converts RGB image (Bayer pattern) to an ARFF file.

@param R/G/B image files
@param bayer pattern
@return one 'output.arff'

Rotate

Rotates an image in left or right direction by 90 degree.

@param input image
@param direction [l,r]
@return image

RoundMaskRatio

Determine the difference in the brightness of a inner circle to the area of a outer ring around that circle. Function is related to BlSpotMatcher in IAP.

@param input image
@param target File
@param outer circle radius
@param dark or bright background
@return contrast image

SideSmooth

Smoothes the left and right-hand side borders of a single object within the image. Uses a polynom to fit a curve and reconstructs the image object with the smoothed-out side borders.

@param binary image
@return binary image

Split

Splits leaves within image (objects which all reach a certain top-position and which at the same time reach a certain lower-end position. Objects which don't 'touch' these virtual top and lower borders, which are of less height, are removed. Then each object which full-fills these criteria, a new image with only that specific object is created.

@input image file
@return file_1.png, file_2.png, file_3.png, ... - objects separated

SplitHistBased

Splits leaves within image (objects which all reach a certain top-position and which at the same time reach a certain lower-end position. Objects which don't 'touch' these virtual top and lower borders, which are of less height, are removed. Then each object which full-fills these criteria, a new image with only that specific object is created.

@param image file
@return file_1.png, file_2.png, file_3.png - objects separated, sorted by x-position of the segment centers

SplitRGB

Splits RGB image into separate channel images.

@return channel0.png, channel1.png, channel2.png or .tif files.
@param image - input image

SplitTiffStackToImages

Splits input Tiff-stack into separate images (e.g. useful for hyper-spectral datasets).

@param tif stack
@return separate images, channel_1.tif, channel_2.tif, ...

ThresholdGTforFGBG

Apply threshold to ground truth (GT) images to generate binary foreground- and background-masks, and the corresponding ARFF files for classifier training.

@param colored ground truth image
@return mask_1.png, mask_2.png (forground and background mask images), and according mask_1.arff, mask_2.arff files!

TransformCSV

Reads the CSV result file and transforms some rows to columns and renames infection color ids to human-readable traits. Cleans-up some content of the column entries, fills empty with 0.0.

@param csv file
@return csv file