Digital histology

Spine Analyzer allows to semi-automatically segment dendritic spines in 3D+t images and to measure their volumes and the intensities of the signal within in different channels over time.

TissUUmaps is a browser-based tool for fast visualization and exploration of millions of data points overlaying a tissue sample. TissUUmaps can be used as a web service or locally in your computer, and allows users to share regions of interest and local statistics.

The tool exports rectangular regions, defined with the NDP.view 2 software (hammatsu) from the highest resolution version of the ndpi-images and saves them as tif-files.

Click the button and select the input folder. The input folder must contain pairs of ndpi and ndpa files. The regions will be exported to a subfolder of the input folder names zones.

Histology Topography Cytometry Analysis Toolbox (histoCAT) is a package to visualize and analyse multiplexed image cytometry data interactively. It can also export data in.fcs data for further analysis using  a specialized cytometry sofwtare such as Flowjo. 

It can be run as a compiled standalone or from matlab.

Using a Hamamatsu slide scanner such as the NanoZoomer, you may end up with NDPI files that can't always be directly open in standard image analysis software such as ImageJ. NDPITools is a collection of software that can convert NDPI files to standard TIFF files, possibly cutting them into smaller JPEG or TIFF pieces that will better fit into your computer's memory. It comes with a bundle of plugins for ImageJ which enable the use of the software directly inside ImageJ with point-and-click.

This small plugin demonstrates the use of OpenSlide in java: it  will extract an imageJ roi drawn from the thumbnail of the whole slide image, or the full image at the desired resolution from an hammatsu NDPI file. Note that z stack are not supported by openslide (neitheir ndpiS).

BioImage.IO -- a collaborative effort to bring AI models to the bioimaging community. 

• Integrated with Fiji, ilastik, ImJoy and more
• Try model instantly with BioEngine
• Contribute your models via Github

This is a database of pretrained deep Learning models. 

Dragonfly is a software platform for the intuitive inspection of multi-scale multi-modality image data. Its user-friendly experience translates into powerful quantitative findings with high-impact visuals, driven by nuanced easy-to-learn controls.

For segmentation: It provides an engine fior machine Learning, Watershed and superpixel methods, support histological data .

It offers a 3D viewer, and python scripting capacities .

It is free for reserach use, but not for commercial usage.

CRImage a package to classify cells and calculate tumour cellularity

CRImage provides functionality to process and analyze images, in particular to classify cells in biological images. Furthermore, in the context of tumor images, it provides functionality to calculate tumour cellularity.

AssayScope is an intuitive application dedicated to large scale image processing and data analysis. It is meant for histology, cell culture (2D, 3D, 2D+t) and phenotypic analysis. 

A deep-learning solution for stain color normalization in digital histology images

ZEN and APEER – Open Ecosystem for integrated Machine-Learning Workflows

Open ecosystem for integrated machine-learning workflows to train and use machine-learning models for image processing and image analysis inside the ZEN software or on the APEER cloud-based platform

Highlights ZEN

• Simple User Interface for Labeling and Training
• Engineered Features Sets and Deep Feature Extraction + Random Forrest for Semantic Segmentation
• Object Classification workflows
• Probability Thresholds and Conditional Random Fields
• Import your own trained models as *.czann files (see: czmodel · PyPI)
• Import "AIModel Containes" from arivis AI for advanced Instance Segmentation
• Integration into ZEN Measurement Framework
• Support for Multi-dimensional Datasets and Tile Images
• open and standardized format to store trained models

ZEN Intellesis Segmentation

ZEN Intellesis - Pretrained Networks

Intellesis Object Classification

Highlights Aarivis AI

• Web-based tool to label datasets to train Deep Neural Networks
• Fully automated hyper-parameter tuning
• Export of trained models for semantic segmentation and AIModelContainer for Instance Segmentation

Annotation Tool

Orbit Image Analysis is a free open source software with the focus to quantify big images like whole slide scans.

It can connect to image servers, e.g. Omero.
Analysis can be done on your local computer or via scaleout functionality in a distrubuted computing environment like a Spark cluster.

Sophisticated image analysis algorithms incl. tissue quantification using machine learning, object segmentation and classification are build in. In addition a versatile API allows you to enhance Orbit and to run your own scripts.

The skin tools measure the thickness of the epidermis and the interdigitation index.

The input images are masks that represent the epidermis and that have been created from images of stained histological sections. The mask must touch the left and right border of the image. The dermal-epidermal border must be on the lower site of the image. The interdigitation index can be measured for one or more segments per image. As a measure of the thickness of the epidermis the lengths of a number of random line segments are measured. The line segments start at the lower border, are perpendicular to the lower border and end at the opposite border of the mask.

See installation Instructions on the website.

The Adipocytes Tools help to analyze fat cells in images from histological section. This is a rather general cell segmentation approach. It can be adapted to different situations via the parameters. This means that you have to find the right parameters for your application.

Sample Image: [0178_x5_3.tif](http://dev.mri.cnrs.fr/attachments/190/0178_x5_3.tif)

Analysis of adipocyte number and size. The original code and example images supposed to be discovered at http://webspace.buckingham.ac.uk/klanglands/ but currently the webpage is missing the code and sample images.

The ultimate goal of the NET framework is to make images of networks processable by computers. Therefore we want to have a pixel based image as input, as output we want a representation of the network visible in the image that retains as much information about the original network as possible. NET achives this by first segmenting the image and then vectorizing the network and then extracting information. The information we extract is

• First and foremost the graph of the network. We find the crossings (nodes) and connections between crossings (edges) and therefore extract information about the neighborhood relations, the topology of the network.
• We also extract the coordinates of all nodes which enables us to embed them into space. We therefore extract information about the geometry of the network.
• Last but not least we track the radii of the edges in the extraction process. Therefore every edge has a radius which can be identified with its conductivity.

In the following we will first provide detailed instructions on how to install NET on several platforms. Then we describe the functionality and options of each of the four scripts that make up the NET framework.

Kappa is a Fiji plugin for Curvature Analysis.

It allows a user to measure curvature in images in a convenient way. You can trace an initial shape with a B-Spline curve in just a few clicks and then fit that curve to image data with a minimization algorithm. It’s fast and robust.

QuantCenter is the framework for 3DHISTECH image analysis applications. with the goal of helping the pathologists to diagnose in an easier way. QuantCenter, is optimized for whole slide quantification. It has a linkable algorithm concept that tries to provide an easy-to-use and logical workflow. The user has different quantification modules that he or she could link one after other to fine-tune or to speed up the analysis.

QuPath is open source software for Quantitative Pathology. QuPath has been developed as a research tool at Queen's University Belfast.

This is the "prediction step" of the Pyxit segmentation model builder. It is a learnable segmentation algorithm based on ground-truth images and segmentation mask. It learns a multiple output pixel classification algorithm. It downloads from Cytomine-Core annotation images+alphamasks from project(s), build a segmentation (pixel classifier) model which is saved locally. Typical application: tumor detection in tissues in histology slides. 

This is a learnable segmentation algorithm based on ground-truth images and segmentation mask. It learns a multiple output pixel classification algorithm. It downloads from Cytomine-Core annotation images+alphamasks from project(s), build a segmentation (pixel classifier) model which is saved locally. Typical application: tumor detection in tissues in histology slides. It is based on "Fast Multi-Class Image Annotation with Random Subwindows and Multiple Output Randomized Trees" http://orbi.ulg.ac.be/handle/2268/12205 and was used in "A hybrid human-computer approach for large-scale image-based measurements using web services and machine learning" http://orbi.ulg.ac.be/handle/2268/162084?locale=en

Cytomine is a rich internet application using modern web and distributed technologies (Grails, HTML/CSS/Javascript, Docker), databases (spatial SQL and NoSQL), and machine learning (tree-based approaches with random subwindows) to foster active and distributed collaboration and ease large-scale image exploitation.

It provides remote and collaborative principles, rely on data models that allow to easily organize and semantically annotate imaging datasets in a standardized way (using user-defined ontologies associated to regions of interest), efficiently support high-resolution multi-gigapixel images (incl. major digital scanner image formats), and provide mechanisms to readily proofread and share image quantifications produced by any image recognition algorithms.

By emphasizing collaborative principles, the aim of Cytomine is to accelerate scientific progress and to significantly promote image data accessibility and reusability. Cytomine allows to break common practices in this domain where imaging datasets, quantification results, and associated knowledge are still often stored and analyzed within the restricted circle of a specific laboratory.

This software is e.g. being used by life scientists in to help them better evaluate drug treatments or understand biological processes directly from whole-slide tissue images (digital histology), by pathologists to share and ease their diagnosis, and by teachers and students for pathology training purposes. It is also used in various microscopy applications.

Cytomine can be used as a stand-alone application (e.g. on a laptop) or on larger servers for collaborative works.

Cytomine implements object classification, image segmentation, content-based image retrieval, object counting, and interest point detection algorithms using machine learning.

Adiposoft is an automated Open Source software for the analysis of adipose tissue cellularity in histological sections.

Example data can be found on the plugin description page in ImageJ wiki (download link). There is also a link to a MATLAB version of the workflow.

CellDetector can detect cells (or other objects) in microscopy images such as histopathology, fluorescence, phase contrast, bright field, etc. It uses a machine learning-based method where a cell model is learned from simple dot annotations on a few images for training and predict on test sets. The installation requires some efforts but the instruction is well explained. Training parameters should be tuned for different datasets, but the default settings could be a good starting point.

This article Baslat et al. presents a method to compute Lymphatic Vessel Density on an image of the whole slide (a workflow documented as text).

Vessels are obtained with a Maximum Entropy Thresholding applied on the excess Red channel (2 times the red values minus blue+green value). Stroma tissue is obtained with a Moment Preserving Thresholding on the blue channel.

The overall colors seen in H&E stained slides can vary widely, influenced by factors such as the precise stains and scanner used. This MATLAB function implements the color normalization strategy described in Macenko et al (2009) in order to match stain colors in an image more closely to 'reference' stains. This may help when comparing images visually, or when applying an automated analysis algorithm.

The function may also be useful to understand the functioning of the color deconvolution described in Ruifork and Johnston (2001).

Quantification of HER2 immunohistochemistry.

ImmunoMembrane is an ImageJ plugin for assessing HER2 immunohistochemistry, described in [bib]2472[/bib]. It is important to read the URL documentation and original paper to understand how to use the plugin appropriately.

There is web service available. Users can upload image data to process them and get cell membrane to be segmented: Web ImmunoMembrane

Note also that the pixel size is not read automatically from the image, but rather the source image scale should be entered into the dialog box - and the image rescaled accordingly prior to analysis. This scale value is the inverse of the value normally found for pixel width and pixel height under Image -> Properties... (i.e. pixel width & height are given in microns per pixel; the dialog box asks for pixels per micron).

SlideToolkit is a collection of command-line tools to assist with the automated histology analysis of whole-slide images. The publication linked in the "reference" details the actual workflow. 

This includes tools to organize the data, perform tiling and subsequent batch processing of the generated tiles in a cell profiler pipeline. All the tools are designed to run on a single PC or on a HPC system. The scripts in the toolkit are on github under MIT licence.

>OpenSlide is a C library that provides a simple interface to read whole-slide images (also known as virtual slides). Python and Java bindings are also available. The Python binding includes a Deep Zoom generator and a simple web-based viewer. The Java binding includes a simple image viewer.  

Analyzing ER, PR, and Ki-67 immunohistochemistry

ImmunoRatio is an ImageJ plugin to quantify haematoxylin and DAB-stained tissue sections by measuring the percentage of positively stained nuclear area (labeling index), described in [bib]2452[/bib].

Notes for use:

• It is important to read the URL instructions and original paper to understand what is being measured. In particular, the primary measurement made is percentage of the total nuclear area, not the percentage of detected nuclei (the latter being the more common method of assessing e.g. Ki67). This may be further modified by the Result correction equation.
• Ultimately ImmunoRatio relies on thresholding (color deconvolved [bib]2451[/bib]) images to define 'nucleus' vs 'non-nucleus' regions according to staining intensity. Therefore dark artefacts, such as tissue folds, are likely to cause errors.
• The pixel size is not read automatically from the image, but rather the source image scale should be entered into the dialog box - and the image rescaled accordingly prior to analysis. This scale value is the inverse of the value normally found for pixel width and pixel height under Image -> Properties... (i.e. pixel width & height are given in microns per pixel; the dialog box asks for pixels per micron).

Web application: ImmunoRatio

Example Image: Sample ImmunoRatio results

References

1. [2452] Tuominen VJ, Ruotoistenmäki S, Viitanen A, Jumppanen M, Isola J.  2010.  ImmunoRatio: a publicly available web application for quantitative image analysis of estrogen receptor (ER), progesterone receptor (PR), and Ki-67.. Breast Cancer Res. 12(4):R56.
2. [2451] Ruifrok AC, Johnston DA.  2001.  Quantification of histochemical staining by color deconvolution.. Anal Quant Cytol Histol. 23(4):291-9.

This macro batch processes all the 2D images (tif and jpg files) located in a user defined folder by calling Fiji Weka trainable segmentation to classify each pixel, and reports the areas of each class in a human readable results table. The classifier to be applied to each image should be previously trained on a representative image by an expert and exported to file (Save classifier) into the image folder to be processed.

[no download link, this description itself explains the steps to quantify staining in tissue sections] The Color Deconvolution plugin for ImageJ can be used to digitally separate up to three stains from brightfield images, after which standard ImageJ commands can be used. The algorithm is described in Ruifork and Johnston (2001). **However**, it is **very** important to take into consideration the caveats on the linked URL. In particular, note that: - Stain colors depend on numerous factors, such as the precise stains and scanner; therefore, the 'default' stain vectors (used to define the colors) are unlikely to be optimal and may be very inaccurate. See the URL instructions for how to create new stain vectors. - Pixel values should be interpreted with extreme caution; in particular, note the warning regarding 'brown' staining that *attempting to quantify DAB intensity using this plugin is not a good idea*. Note, the pixel values provided by this plugin are 8-bit and **not** equivalent to 'optical densities' frequently presented in the literature. Color deconvolution is particularly helpful in separating stains so that stained regions can be detected (e.g. by setting a threshold), and then the number or areas of stained structures may be quantified. Two potential approaches would be: 1. If one measurement should be made for the entire image: - *Image > Adjust > Threshold...* - *Edit > Selection > Create Selection* - *Analyze > Measure* 2. If distinct structures should be measured: - *Image > Adjust > Threshold...* - *Analyze > Analyze Particles...*

An ImageJ macro for calculating empty surfaces on histological slices (ex: tubules in a kidney).

ilastik is a simple, user-friendly tool for interactive image classification, segmentation and analysis. It is built as a modular software framework, which currently has workflows for automated (supervised) pixel- and object-level classification, automated and semi-automated object tracking, semi-automated segmentation and object counting without detection. Most analysis operations are performed lazily, which enables targeted interactive processing of data subvolumes, followed by complete volume analysis in offline batch mode. Using it requires no experience in image processing.

ilastik (the image learning, analysis, and segmentation toolkit) provides non-experts with a menu of pre-built image analysis workflows. ilastik handles data of up to five dimensions (time, 3D space, and spectral dimension). Its workflows provide an interactive experience to give the user immediate feedback on the quality of the results yielded by her chosen parameters and/or labelings.

The most commonly used workflow is pixel classification, which requires very little parameter tuning and instead offers a machine learning technique for segmenting an image based on local image features computed for each pixel.

Other workflows include:

Object classification: Similar to pixel classification, but classifies previously segmented objects by object characteristics in a subsequent step

Autocontext: This workflow improves the pixel classification workflow by running it in multiple stages and showing each pixel the results of the previous stage.

Carving: Semi-automated segmentation of 3D objects (e.g. neurons) based on user-provided seeds

Manual Tracking: Semi-automated cell tracking of 2D+time or 3D+time images based on manual annotations

Automated tracking: Fully-automated cell tracking of 2D+time or 3D+time images with some parameter tuning

Density Counting: Learned cell population counting based on interactively provided user annotation

Strengths: interactive, simple interface (for non-experts), few parameters, larger-than-RAM data, multi-dimensional data (time, 3D space, channel), headless operation, batch mode, parallelized computation, open source

Weaknesses: Pre-built workflows (not reconfigurable), no plugin system, visualization sometimes buggy, must import 3D data to HDF5, tracking requires an external CPLEX installation

Supported Formats: hdf5, tiff, jpeg, png, bmp, pnm, gif, hdr, exr, sif

MembraneQuant performs automatic evaluation on IHC membrane stainings (HER2, EGFR etc).

Using color deconvolution, MembraneQuant detects cell membrane and measures staining intensity on the chromogen channel. This way it is possible to calibrate the software to the actual stain protocol in the pathology lab. The algorithm categorizes the detected membrane to weak positive, medium positive and strong positive classes.

This software has an IVD certification for HER2 quantification.

When opening the Pannoramic Viewer you see all of your virtual slides in thumbnail view. Selecting one (or up to 10 at a time) the slide gets under the virtual objective of the virtual microscope. Here you can move and change the magnification of the slide quickly and easily using the mouse. Emphasizing 'quickly' is important considering the fact that the size of an average virtual slide can easily be more than 1 GB.

Main characteristics:

• Seamless zooming and moving of the virtual slide
• Bookmarking (annotating) on the spot, i.e. defining the specific part of the sample by drawing; finding and reading of previously made bookmarks
• Easy and precise measurements
• Real-time changing of brightness, contrast and color bias
• Fluorescent slide handling, separate channel view & pseudo-colorization
• Slide uploading and downloading for teleconsultation
• Synchronized viewing (moving and zooming) of multiple slides for comparison purposes
• Publication quality image capture of displayed areas (.JPG, .BMP, .TIFF)
• TIFF, MIRAX slide and Meta-XML export for Carl Zeiss AxioVision™ compatibility
• Scanmap export for rescanning existing digital slides
• Easily expandable functionality via the software modules

NuclearQuant is a QuantCenter module. It is designed for cell nuclei detection and quantification of IHC stained samples. NuclearQuant measures several morphological features besides stain intensity. The cell nuclei classification and the final score are calculated by the intensity score and the proportion score.

DensitoQuant is a simple and fast, yet effective tool for IHC measurements. It measures the density of immunostain on the digital slides by distributing pixels to negative and 3 grades of positive classes by their RGB values. DensitoQuant is especially suitable for quick TMA evaluation. Analyzing a whole digital slide takes only a couple of minutes.

This plugin is bundled with Fiji. For installation in ImageJ1, download from the link below and manually install the class file. 

Quote:

The colour deconvolution plugin (java and class files) for ImageJ and Fiji implements stain separation using Ruifrok and Johnston's method described in [1]. The code is based on a NIH Image macro kindly provided by A.C. Ruifrok.
The plugin assumes images generated by colour subtraction (i.e. light-absorbing dyes such as those used in bright field histology or ink on printed paper). However, the dyes should not be neutral grey (most histological stains are not so).
If you intend to work with this plugin, it is important to read the original paper to understand how new vectors are determined and how the procedure works.
The plugin works correctly when the background is neutral (white to grey), so background subtraction with colour correction must be applied to the images before processing.
The plugin provides a number of "built in" stain vectors some of which were determined experimentally in our lab (marked in the source with GL), but you should determine your own vectors to achieve an accurate stain separation, depending on the stains and methods you use. See the note below.
The built-in vectors are :

• Haematoxylin and Eosin (H&E) x2
• Haematoxylin and DAB (H DAB)
• Feulgen Light Green
• Giemsa
• Fast Red, Fast Blue and DAB
• Methyl green and DAB
• Haematoxylin, Eosin and DAB (H&E DAB)
• Haematoxylin and AEC (H AEC)
• Azan-Mallory
• Masson Trichrome
• Alcian blue & Haematoxylin
• Haematoxylin and Periodic Acid - Schiff (PAS)
• RGB subtractive
• CMY subtractive
• User values entered by hand
• Values interactively determined from rectangular ROIs