Biomedical Imaging

Extract more from images
to drive Drug Discovery with AI

Throughout many years, information contained in images was underappreciated and unused due to the limitations of human capabilities. However, thanks to recent breakthroughs and our own expertise in Computer Vision, we are able to guide you through the effective analysis of vast amounts of images, which are an especially rich source of information.

 

Our Technology Platform for Biomedical Imaging, provides you with reliable and interpretable AI models, for both clinical and preclinical usage.

 

Overcome your challenges in image analysis to unlock insights from your experiments

Biomedical images are helpful sources for scientists and practitioners in drawing significant hypotheses, exemplifying approaches, and describing experimental results. It is one of the highly complex and unaccomplished tasks scientists and physicians face today when it comes to implementing a system, which can help them collect, curate, annotate and validate information distributed among biological images.

There has been a significant increase in the amount of heterogeneous biomedical image generation, which results in a need for bioimaging platforms for feature extraction and analysis. This could give an advantage in implementing effective information retrieval systems.

It’s not easy to leverage full information from images. We provide Computer Vision solutions to many clients that are struggling with various difficulties. Some of them were related to:

Are you facing similar challenges?

Large input filesImages streaming fromdifferent devices andexperimentsVarious resolutionsandt quality issuesSmall training setsOr large numberof missing labels
Discovery and clinical phase

Digital histopathology

Accelerated learning to reveal more from your biopsy results

Our platform enables you to process whole-slide biopsy images in a scalable manner, without any loss of information contained in the image and with instant quality check and image standardization. Moreover, it  highlights  areas that contain pathological changes and performs automatic segmentation of the tissue. This process can be applied to various scoring systems, such as the Nancy and Geboes Histological score (for inflammatory bowel disease data) or the Gleason score (for prostate cancer data).

Problem
definition

Obszar roboczy 1Histology images are one of the most difficult to analyze automatically.Smalltraining setQuality issuesof acquired imagesMislabelingDifferent sizesof biopsiesLarge input files - up to 4GBper imageOnly smallregions relevantThis is due to multiple factors, such as:

Value we deliver

  • leveraging expertise of histopathologists 
  • meaningful and interpretable predictions 
  • proven performance comparable to an expert pathologist 
  • automatic detection of quality flaws 
  • more reliable and precise diagnostics
Discovery phase

High content screening

Image based virtual screening to increase the chance for success of your drug candidate

Our platform solves the challenges of a vast number of image analyses by providing modular deep learning pipelines that generate image based compound descriptors, predict selected molecular properties and activities towards selected targets and automatically retrain to fit newly generated data. It also identifies the drug and drug targets that modulate specific phenotypes. High Content Phenotypic Screening aims to identify morphological patterns in cell populations and to probe the causes and cures for various diseases.

Problem
definition

Various imageresolutionsImages coming from various imaging devices and experimentsImages acquiredat differentcompoundconcentrationsLarge numberof missing labelsVarious numberof channels

Value we deliver

  • acceleration of the drug discovery process
  • improved accuracy
  • meaningful and interpretable prediction
  • automatic detection of quality flaws
Discovery phase

Microarrays

An automated and more reliable process for your custom microarray analysis

Microarray technology has gained momentum in the drug and biomarker discovery fields. Extracting information from microarrays requires precision and reliability, and our microarray analysis system delivers it all. It performs quality checks and applies image correction techniques, such as distortion correction or flat-field correction. It also finds the grid of features on the array and extracts their pixel intensities, creating automatic segmentation of regions of interest. Additionally, the system can be deployed in the cloud and easily integrated into the already existing analytic pipelines.

Problem
definition

Custom-designeddense microarrayassaysLimited availabilityof software to enable acquisition, qualitycheck, processing andautomatic analysisDifferentmicroarraydesignsDifferentmicroarray imageresolutions Variable quality andbrightness of image patches

Value we deliver

  • fully automated data acquisition
  • integrated microarray image quality check
  • more reliable process of microarray usage
  • improved statistical properties of the platform
  • FDA-compliant process & documentation

Technology Platform for Biomedical Imaging

The Ardigen platform is applying various algorithms, deep neural networks and machine learning techniques for analyzing biomedical images to help scientists and physicians understand and predict the behavior of complex biological systems.

We strive to improve accuracy and efficiency of diagnostic and therapeutic procedures by utilizing our biomedical computing competences such as: digital image acquisition, processing, quantitative analysis and artificial intelligence in biomedicine; as well as human-computer-interaction and user experience engineering.

ReliableInterpretableAutomaticPreprocessing& quality checkAutomaticre trainingUser feedback /active learningInterpretability& visualizationModel diagnostics Inference & prediction AutomaticTrainingModel versioningsystem & reproducibilityUser outcomeExperiment resultsDataUser feedback + data

Our science

We are driven by science and collaborations!

Since 2015, we’ve had the opportunity to execute a variety of exciting and complex projects, both for our clients and internally.
A high degree of trust and confidentiality is the foundation for our business. That’s why we are always very excited to share the results of our scientific projects.
We publish and present materials, such as journal publications, posters, case studies and technical notes whenever possible, through which we aim to show the scope of our scientific team’s experience.

On demand webinars

From sequences to properties of proteins – Speed up your drug discovery process with AI
27 April 2021 From sequences to properties of proteins – Speed up your drug discovery process with AI

    Digital pathology – AI for biopsy evaluation
    19 March 2021 Digital pathology – AI for biopsy evaluation

      A single pixel is worth its weight in gold. Automatic microarray analysis platform
      17 March 2021 A single pixel is worth its weight in gold. Automatic microarray analysis platform

        Meet us at upcoming events

        • Coming soon
          15 - 17 June
          digital
          Pharma R&D Week Virtual
        • 14 - 18 June
          digital
          PMWC 2021 COVID Virtual
        • 23 - 26 June
          digital
          SLAS Europe 2021
        • 9-10 June
          digital
          Precision in Drug Discovery & Preclinical Virtual Summit

        Contact us

        Are you interested in Ardigen TCRact or ArdImmune Vax Platforms based solutions?

          • https://ardigen.com/wp-content/uploads/2021/05/michal_warchol-min-1.png Michał Warchoł, PhD General Director of Biomedical Imaging Unit
          • https://ardigen.com/wp-content/uploads/2021/05/circle-cropped-1.png Katarzyna Karpińska Business Development Manager
          Go up