Analysis Pipeline - 2D/3D/4D Image Analysis
arivis Vision4D is known for its capability to segment and track objects in data sets of any size, from mega- to terabytes. To achieve this, arivis Vision4D provides a robust and fast tool for processing of image data acquired from long-term time lapse experiments.
The Analysis Pipeline is a powerful solution with an intuitive interface for filtering and analyzing image data. The pipeline approach enables novice users to quickly run a pre-existing or newly built analysis workflows, while the image analysis expert is still empowered to create complex analysis routines as required.
Users can freely combine different operators for filtering, segmentation and other analysis tasks in a simple and clear pipeline. Custom-defined analysis pipelines are processed interactively with many operations providing preview options so users can analyze large datasets confident in the quality of the results. Analysis results can be reviewed in both 2D and 3D, or both simultaneously, with interactive charts and colouring options to facilitate interpretation. The interactive charts allow for fast and easy selection of objects or population groups based on one or two of a wide range of features that can be collected from segmented objects.
Results tables are easily customized to include a range of built in and custom features than can be readily exported to XLS or CSV documents for further statistical analysis if needed.
Charts and segmented images can be easily shared as graphics for presentations or publication through the one-click snapshot feature.
To facilitate the creation of quality pipelines while maintaining interactivity even when dealing with very large images, arivis Vision4D makes it really easy to work on subsections of a dataset without the need to crop and or duplicate the data. The region selection can be based on built in presets for the current view, plane or time point, or custom selection of ranges of times and planes within specific bounds that can be defined manually or taken from previous segmentation.
Multiple pipelines can be run on the same datasets independently or in conjunction with other pipelines to permit more complex analysis or extract different features types from the same dataset.
Define your segmentation workflow by adding image processing and analysis operators to your pipeline
- Apply voxel operations for image enhancements and editing. These include:
- Masking based on thresholds or previous segmentation
- Advanced image filters for denoising, membrane enhancements, morphological operation etc
- Advanced image processing images in MATLAB thanks to MATLAB COM integration, or through the trainable filter
- Advanced segment generation operation that work on both 2D or 3D images from a range of imaging modalities (brightfield, fluorescence, EM, CT and many more)
- Threshold based segmentation on monochrome and colour images using manual or automatic thresholds
- Advanced object detections algorithms
- Integrated Machine Learning
- Advanced object classification tools
- Easy storage and export options built straight into the pipeline for easy integration in batch processing
- Store analysis results directly in the SIS file for easy access and review
- Export features directly to XLS files for further statistical analysis in statistical packages
- Export meshes to OBJ files for easy integration in CAD packages or other 3D viewers
- Advanced pipeline handling
- Save and export your pipelines to use them on other images – create pipelines that others can use and modify
- Select pipelines in the batch process to automate the analysis of multitudes of datasets
- Import pipelines into arivis WebView for processing directly on the online server
Flexible pipeline execution
- Combine different operators and filter methods (noise reduction, blur filters, contour filters, edge enhancement, etc.)
- Preview the results of most operations directly on the image so operations can be run with the confidence that the results will be appropriate
- Easily progress through all or part of the pipeline with easy rollback to test and adjust operational parameters
- Immediate result visualization and a live analysis report allows an evaluation and correction of parameters
- The Blob Finder analysis operator is ideal for segmenting cells or any other type of rounded cell organelles in a noisy image in an easy-to-use 3-step process
Viewing and Exporting Analysis Results
Along with preview options for many operations, the ability to immediately view results on the image makes it easy to evaluate the quality of the analysis. Object outlines are easily discernible in both 2D and 4D views, with advanced viewing options including custom colouring and transparency making it easy to contextualize the results. Image views, including objects, can easily be exported with 1-click snapshots or high resolution renders.
Along with visualization of the objects on the image, advanced data tables allow for precise interaction with the data through the Objects Window. Double-click an object in the Objects Window and immediately see it on the image, or select an object in the image to see it highlighted in the data table. The Objects Window is highly customizable, allowing users to view only relevant object features, with advanced graphing options based on time line progression, histogram distribution or even 2D distributions for easy population analysis. The Objects Window can be shown automatically after each segment operation for quick and easy review, or loaded on demand for an uncluttered workflow. And of course, the content of the Objects Window can easily be exported to statistical analysis packages along with other export options.
Proof reading and editing made easy
Despite using some of the most advanced segmentation tools, object detection is not always foolproof. However, with arivis Vision4D users have powerful tools to refine automatic segmentations:
- Manual create objects: Draw a polyline around an object on consecutive 2D slices or just a few slices throughout the depth of an object and arivis Vision4D will interpolate the complete object outline in 3D. This new manual segmentation tool is ideally suited to touch screen devices, and combined with arivis Vision4D’s low hardware requirements makes it possible to work even on tablets.
- Magic Wand semi-automatic object detection: Easily detect objects based on simple thresholds within a user defined region in one click. This tool works on both 2D and 3D views with depth awareness for easy interaction.
- Brush tool: Enable manual editing and work on segmented objects in sequential planes using the brush tool to fill in gaps or erase unwanted portion then confirm your changes and the object is updated with its new features recalculated.
Measurements of Objects
The Objects window can show a wide range of measurement for the segmented objects. With the user selecting what measurements to include there’s no cluttered tables with irrelevant measurements. The objects measurements are calculated dynamically and displayed in the Objects Window where they can be easily sorted. Custom object measurements allow the user to expand the characterization beyond what’s included out of the box.
arivis Vision4D offers the ability to measure using a range of measurement types, from 3D polylines to complex volumes and tracks. Each of these object types enable the collection of specific features, but also includes measurements that are independent of object type including:
- Rectangular or elliptical bounds of the object and their characteristics measured in either calibrated units or pixels (width, height, start/end position…)
- Centroids of objects based on geometry or bounding box
- Temporal information relating to the position of an object within the dataset
- much more
Along with measurements common to all object types arivis Vision4D also includes measurements specific to the types of objects in question, including:
- 3D volumes
- Intensities (mean, minimum, maximum…)
- Volume (with/without holes, calibrated/uncalibrated)
- Surface area
- Length (total, segments)
- Custom object features: Take any measurement of an object and create more complex measurements including:
- Ratio of one measurement against another
- Nearest object distance
- Group statistics
- Combinations of complex measurements
Flexible statistical and graphing tools allow users to get more from their data through grouping and without the need to export measurement tables to Excel.
- Sort the Objects Window with one click according to any measurement
- Change the units to make easier to read the results
- Colour objects according to measurement value for easy pinpointing of outliers
- Plot histograms of distribution and use 2D graphs to easily identify subpopulations
- Combine objects into groups to quickly get more targeted statistics
- Analyse parent/child relationships using segment colocalization
Image analysis that fits your data
The advanced image analysis tools, combined with the flexible statistical analysis make image analysis fit your data. Freely specify relationships between different components to obtain more meaningful results.
Segments and tracks can be grouped according to feature ranges or custom selection.