From Image
to Insight

Advancing Biomedical
Research with BIAS

BIAS processes and interprets high-content and tissue microscopy imaging. It segments, categorizes, annotates, and measures regions, down to the single-cell level.

Why BIAS?

Fully Customizable BIAS is a powerful AI-based image-processing platform offering dozens of algorithm options for every stage of your workflow. Each algorithm is highly customizable to meet your needs, giving you the flexibility to create and tailor unique image-processing pipelines with ease.

Automated Workflows. Save your pipelines and apply them to new datasets for streamlined, automated processing of large data volumes. With BIAS, you can deploy and run pipelines without user interaction, enabling batch processing and maximizing efficiency.

Train your own classifier interactively for specific phenotypes, extract detailed statistics, and validate accuracy with built-in tools. Once trained, models can be reused on new datasets. No need for retraining each time.

What is new?

Advanced Deep Learning for Phenotype Classification

Tackle even more complex phenotype classification challenges with our powerful deep learning-based classifier. Train and validate models interactively, just like our traditional classifiers, but with the added precision and flexibility of deep learning.

Seamless Pipeline Integration for 3rd-Party Programs

Easily integrate your own or third-party processing algorithms into BIAS pipelines with our new Execute module. Using a simple command-line interface, you can seamlessly integrate multiple algorithms into any stage of the pipeline without restrictions.

Simple Project Import & Export

Easily share and distribute your projects and pipelines with colleagues using BIAS’s new import and export functionality. Save time and effort by seamlessly transferring work between teams and streamlining collaboration.

Features

BIAS offers various features. Users can customize cutting-edge algorithms, automate workflows, and train classifiers interactively. Find out more about all the BIAS features below.

1

Image Pre-Processing

BIAS offers multiple pre-processing options for microscopy images, including intensity normalization, flat-field correction, resampling, and retiling.

2

Detection and segmentation

BIAS provides multiple methods for detecting and segmenting cells and subcellular structures in both 2D and 3D, including spot detection, classical segmentation, deep learning-based models, and mask post-processing for enhanced results.

3

Feature Extraction

BIAS enables the extraction of features from segmented objects across all sub-structures and channels. Features include shape (size, circularity, etc.), texture, signal/color intensity, and micro-environment characteristics.

4

Adaptive Classification

BIAS offers both unsupervised and supervised machine learning for classifying segmented objects based on their features. The supervised module also supports image-based deep-learning models that can be trained to classify different phenotypes. The number of phenotype classes isn't fixed - you can freely add or remove as needed, even during the active

learning process when discovering new phenotypes. Models can be trained and validated on the fly, with various self-validation methods available (such as k-fold cross-validation and leave-one-sample-out cross-validation) to reliably estimate the expected accuracy of your classification models.

5

Statistics, Analysis & Filtering

BIAS enables in-depth analysis of extracted features and classification data, allowing object-level or group assessments. Count, visualize, aggregate, and filter objects based on custom criteria. All data is stored in an internal SQL database, with options for manual queries, a query wizard, or predefined templates for streamlined analysis.

6

Export for Micro-Dissection

BIAS enables the export of selected cells or regions to supported laser micro-dissection systems. Contour export modules automate transformation between microscopes and can transfer optimized cutting paths for thousands of objects. The cutting path and object order can be optimized, and the selection of collection caps and plates can be automated on selected microscopes.

Sub-cellular structures can be independently dissected, allowing targeted collection (e.g., nuclei in one collector, cytoplasm in another).

7

Comprehensive Data Export

BIAS supports exporting nearly all intermediate and final processed data in standard formats, including JPEG/PNG for images, PNG for masks, and CSV for feature matrices and query results. Exported data can be further processed, extended, or reloaded into BIAS to integrate and visualize, e.g. additional -OMICs features. A provided Python package enables seamless reading and writing of BIAS-compatible data from a Python environment.

8

Project Creation & Pipeline Automation

BIAS enables the design of full image processing pipelines connecting any number of algorithms. Pipelines can be saved as projects, shared across datasets, and executed with all parameters and trained states for automated processing. Headless execution enables seamless batch processing for maximum efficiency.

9

Integration of External Algorithms

With the Execute module, any external executable algorithm, written in any language, can be integrated into your image processing pipeline. Simply follow a basic command-line interface structure for effortless insertion and execution.

Interested in seeing
BIAS in action?

Request a personalized demo and let our team show you how BIAS can be configured for your workflows and goals.
We will demonstrate solutions that solve your biomedical imaging challenges live in action.

Request demo

Frequently asked questions

Data management

expand_more

What kind of microscopes does BIAS support?

BIAS is compatible with various microscopes, such as Perkin-Elmer/Revvity Opera and Operetta, 3DHistech, GE IN Cell, and Leica XLEF, in both 2D and 3D. There are no restrictions on the number of color channels, making it ideal for handling multiplexed data.

BIAS supports a wide range of microscope image formats, including brightfield, fluorescent, and pathological images, either directly or through a converter application.

Additionally, BIAS supports standard image formats like JPEG, PNG, and TIFF. It also allows exporting contours for Laser MicroDissection (LMD) in Leica LMD and Zeiss Palm formats.

How do I import data/files into the system?

BIAS converts all external data into a common internal format to ensure compatibility with its processing algorithms.

To import data, users need to run BIAS_IMPORT, an importer program that can be executed either externally or within the BIAS software. During this process, users will be prompted to specify a working folder where the imported data (and later, the processed data) will be stored.

For better organization, multiple working folders can be used.

How do I export data correctly?

BIAS provides a dedicated Exporters module for exporting internal data formats (such as segmentation masks and feature matrices) into standard file formats like PNG and CSV.

For contour export used in microdissection, specialized Contour Export modules are available.

Additionally, BIAS offers a Python package that allows users to read, write, and process BIAS internal data formats without needing to export them.

Image processing

expand_more

What image processing steps can be performed using BIAS?

BIAS offers a range of image processing capabilities, including:

Image post-processing – Flat field correction, resampling, normalization, and more.
Cell and object detection/segmentation – Identifying and segmenting structures within images.
Feature extraction – Analyzing image data to extract relevant features.
Phenotype classification – Using supervised or unsupervised machine learning algorithms.
Basic statistical analysis and cell selection – Supporting quantitative image analysis.
Contour export – For laser microdissection.
Generic data export – Allowing the export of intermediate data for further processing

What type of cells can BIAS detect and segment? What phenotypes can be classified?

BIAS supports both classical segmentation methods (f.x. auto-thresholding) and modern deep-learning-based models based cellular or subcellular structure segmentation models.

Our generic models can handle a wide range of imaging modalities.
We also offer specialized models tailored for specific tasks.

For phenotype classification, BIAS provides an interactively trainable system that supports both feature-based and image-based algorithms. Users can customize classification models to fit the specific needs of their project.

What kind of experience is needed for using BIAS?

BIAS is designed primarily for biologists and researchers, offering extensive customization options for image analysis and more control over the image analysis pipeline. While experience with similar software is recommended, it is not required.

Support

expand_more

I am getting an error message. What does it mean, and how can I fix it?

If you encounter an error message, please contact us with following details:

The exact error message and error code (if available).
A brief description of when and how the error occurred.

Our technical support team is happy to assist you and we will work on finding a solution.

I found a bug in the software. How do I report it and get it fixed?

Please report any bugs through our technical support page (link). Please provide the following details:

BIAS version in which the error occurred.
Operating system you are using.
Steps to reproduce the issue (if possible).

Your input helps us improve BIAS - thank you for your support!

How do I set up/configure a feature correctly?

BIAS includes a built-in user manual, accessible from the Help menu. The manual provides an overview of BIAS's basic functionality and detailed instructions for configuring and using each module. Please refer to the manual for step-by-step guidance on setting up features correctly.

I am having trouble installing or setting up the software. What should I do?

BIAS is a standalone, portable software, meaning no installation is required. If you received BIAS as a compressed file, simply extract it to a suitable location, and it will be ready to run. If you need further assistance or continue to experience issues, please contact our support team (link).

How do I update the software, and what changes are included in the latest version?

BIAS updates are released as new standalone portable versions regularly.

When a new version is available, our team will notify you directly.‍
New releases are typically compatible with older data and pipelines

If needed, users can keep older versions alongside the new release to ensure 100% compatibility.

What operating systems is BIAS compatible with?

BIAS currently supports Microsoft Windows (versions 10 and 11) and Ubuntu Linux (LTS 20.04 and later) as its main platforms.

From Image to Insight

Data management

expand_more

Image processing

expand_more

Support

expand_more

From Image
to Insight