Pattern recognition and machine learning capabilities enable VIAS to perform advanced particle characterization, classification and measurements.
VIAS is an imaging and analysis software that controls transmission electron microscopes (TEM) and provides semi-automatic image acquisition as well as subsequent analysis and data presentation. The software can control the MiniTEM™ and Vironova HT7800 systems.
VIAS is an intuitive software used to control the microscope, visualize samples, and analyse data with little previous knowledge. VIAS provides a step-to-step guide through the sample exchange process and easy-to-use tools to adjust settings for optimal image acquisition.
The software has four modules (tabs):
Start: organizes imaging projects and allows the user to modify global database resources.
Imaging: provides an interface to view live images and manually capture images from the microscope (see Figure 2).
Analysis: displays acquired images and the results of particle classification, analysis and statistics.
Workflow: provides an easy interface to create, modify and run screening or analysis workflows.
Figure 1 - VIAS is an intuitive software used to simplify the analysis of images acquired using transmission electron microscopy (TEM).
From the imaging tab of the software the user can control the microscope and its peripherals to explore a sample. While exploring a sample manually, the software will provide a live feed showing the area visible to the camera given the current microscope settings. Image acquisition can also be performed automatically using either a screening mode that acquires images at pre-determined positions that are identified by the software or at waypoints that are identified by the user.
Figure 2 - The Imaging module allows the user to see and navigate in the current sample and acquired images. This tab also contains all the settings for the microscope and the camera.
The analysis tab offers a range of tools for automatic or manual analysis of image data. For automatic particle detection and classification, particle information is stored as metadata within the corresponding images. The resulting particle populations, found in single images or aggregated over a repository of images, can be further quantitatively analyzed with regards to various morphological properties such as diameter, length, area, or higher order characteristics such as user defined class or type. These characteristics can be visualized using a selection of plotting tools that can then be stored to file (see Figure 4). For manual annotation, the user can measure and identify objects of interest.
VIAS includes a number of ways to show graphical representations of measurements on particles in the images, including:
All plot types support visualizing the particle data grouped by their class, and, depending on plot type, different measures can be selected.
VIAS also provides powerful tools for performing statistical analyses on detected objects in images. The statistics engine allows users to perform calculations within the software, thus eliminating the need to export data to third-party tools. This in turn ensures that statistical analyses are directly linked to source data, reducing the risk of errors caused by duplicate data sets.
Figure 3 - The Analysis module allows the user to measure, inspect and mark the image with colored text, shapes, arrows or regions.
Figure 4 - An example box plot. Plots can be exported as PNG files for use in reports or presentations.
Figure 5 - Statistical analysis result performed on a set of 74 images acquired at 3 waypoints.
The Workflow module of the VIAS software allows for the creation of customized operations using a modular drag-and-drop interface. The system offers many modules for microscope control, image transformation, particle segmentation & classification. Workflows are used to perform automated image acquisition (screening) and automated image analysis with the purpose of detection and classification of particles within a set of images. Workflows are typically designed and adapted to specific particle types, grid preparations, and imaging conditions.
In addition to traditional image analysis tools, VIAS also includes detection models based on convolutional neural network (CNN) technology. CNN detection uses pre-trained models designed for detection of various objects. Depending on the data the models are trained on, they can be specialized and designed for solving specific detection problems or general and adaptable to many different detection problems. New models can be imported into VIAS to add or refine functionality. Models can only be acquired from Vironova.
Figure 6 - Workflows are based on a modular graphic interface.
VIAS can be deployed in two ways:
1) as a microscope control client and
2) as a standalone analysis client.
The standard installation for microscope control includes the microscope hardware with an integrated workstation PC running VIAS, VIAS database administrator, and SQL Server Express for local data storage. By adding the workstation to a networked domain, data storage can be located on SQL servers on the domain and additional VIAS standalone analysis clients (requiring an additional licenses) can be added.
Use of a networked SQL Server is recommended for optimal performance. Advantages of a network setup include:
Figure 7 shows two possible deployments for microscope control. The first is for the Vironova MiniTEM microscope system and the second is for the Vironova HT7800 microscope system.
Figure 7 - VIAS system topologies for microscope control.