AI-based diagnostic support in medicine with »Pneumo.AI«

»Pneumo.AI« is an AI-supported diagnostic software which supports medical professionals in the accurate identification of pneumonia using intelligent image recognition.

In which areas does the AI application offer the greatest benefit?

The »Pneumo.AI« demonstrator was developed to playfully illustrate the ways AI technologies and medical professionals can work together. The disease pneumonia, better known as lung inflammation, keeps specialized medical specialists busy in many hospitals. Until now, possible diseases had to be identified manually using X-ray images, among other things, to initiate appropriate treatment in good time. Today, image recognition systems can help with diagnosis. This not only saves time but can also prevent misjudgments.

What is pneumonia?

Pneumonia is an acute inflammation of the lower respiratory tract, commonly referred to as lung inflammation. Contrary to current belief, pneumonia is still a serious disease in many regions of the world today. In developing countries in particular, pneumonia is one of the most common causes of illness and death in children under the age of five.

How can AI help with diagnosis?

Artificial intelligence systems can support doctors in making a diagnosis. In this case, computer vision, i.e., machine vision, helps to recognize disease characteristics on chest scans.

What is the future of AI in medicine?

Both doctors and data scientists alike see great potential for AI in medicine. Many hospitals have large amounts of data available which could be used to improve diagnostic support. However, it is important that AI systems are only ever understood as assistance tools and that medical staff always remain in charge of decision-making. Also, medical data is highly sensitive and requires special protection.

What are »Pneumo.AI’s« quality characteristics?

  • Low effort: Since the annotation of medical image datasets usually involves a great deal of effort, it is important to develop data-efficient algorithms to achieve the lowest possible annotation effort.
  • Immediate analysis: The use of AI technologies allows for immediate evaluation of the scan/X-ray image after it has been taken – with no human interaction. This shows the potential to optimize work processes in clinics, for instance by developing a prioritization system. However, it is important that AI always serves as an assistance system for doctors and never makes decisions on its own.  
  • Secure data processing: For sensitive data such as patient data, it is essential that the AI processes used are secure. All data must be stored on German servers or may only be processed locally by medical specialists or in hospitals. 
  • Powerful AI: In the future, AI-based multimodal analysis will also play an important role in the evaluation of medical image data, as doctors will have a wide range of information available to them during the diagnostic process regarding the patient’s health status or the course of the disease.
»Close collaboration between medical experts and data scientists is the most important basis for the use of artificial intelligence in medicine.«
Helen Schneider
Fraunhofer Institute for Intelligent Analysis and Information Systems IAIS

Which AI technology is used in the KI.NRW demonstrator?

Deep Learning

The Pneumo.AI demonstrator is based on deep convolutional neural networks (CNNs), which are particularly well suited for processing large image data sets. This technology can also be transferred to other diseases and use cases.

Informed Machine Learning

In this type of machine learning, available prior knowledge and expert knowledge are integrated into the model to develop more data-efficient algorithms, for instance. For Pneumo.AI, the elements of bilateral symmetry of the lung field were considered within the modeling.

Computer Vision

To ensure a good generalization capability of the trained network, various augmentation techniques are implemented. By rotating and zooming the training image data, the network achieves better performance and overfitting is avoided.

What does the AI demonstrator show?

The »Pneumo.AI« demonstrator shows how AI technologies can support doctors in practices and clinics in their everyday work in the future. It is important to emphasize that artificial intelligence is available to medical professionals as an assistance tool, but that the final decision remains with the human being. The demonstrator also illustrates the great potential of AI in medical image processing.

Request a non-binding consultation with our experts now!

Curious? Click here to go straight to the demonstrator!

Where can I find more information?

AI.Map featuring entries from the medical field

More AI providers, applications, and AI products »made in NRW« with the same AI focus can be found using the filter and search function of the AI.Map, which currently contains more than 1000 entries.

Lecture on Pneumo.AI at the MEDICA 2022 trade fair

At the international medical trade fair MEDICA 2022, KI.NRW and Fraunhofer IAIS gave a presentation on »Artificial intelligence in healthcare using the example of Pneumo.AI«.

SmartHospital: The use of AI in the hospital of the future

The KI.NRW flagship project SmartHospital.NRW aims at developing tools to support hospitals in their digital transformation and in the use of AI. Determine your hospital’s AI maturity level now.

Contact the team of developers

Helen Schneider

Data Scientist – Computer Vision

Fraunhofer IAIS
Schloss Birlinghoven
53757 Sankt Augustin

Phone +49 2241 14-2735

Send email

Dr. Rafet Sifa

Head of Cognitive Business Optimization

Fraunhofer IAIS
Schloss Birlinghoven
53757 Sankt Augustin

Phone +49 2241 14-2405

Send email

“Sustain.AI” – The AI tool for analyzing sustainability reports

“Sustain.AI” – The AI tool for analyzing sustainability reports

Where does the AI application offer the greatest benefit?

Sustainability reports are an important part of a company’s information policy. They provide interested members of the public with information about the organization’s activities and performance regarding sustainable development. Since 2017, all listed companies with 500 or more employees have been required to publish such reports. In doing so, they follow the CSR (Corporate Social Responsibility) directive. The main aim of this directive is to increase transparency regarding the environmental and social aspects of companies in the EU. This includes information on environmental, social and employee issues as well as the protection of human rights and the fight against corruption.

Like annual reports, sustainability reports are also used as a basis for important purchasing or investment decisions. However, the necessary identification of all relevant criteria and information usually requires a great deal of time and effort. With the tightening of the CSR Directive implemented by the EU in 2023, which will extend reporting requirements to other aspects and a larger group of companies, this work becomes even more complex – especially if reports are evaluated manually.

The AI-based tool Sustain.AI makes this work easier. Thanks to machine text recognition, sustainability reports can now be analyzed in a highly efficient and structured way. The technology behind Sustain.AI is particularly aimed at auditors and controllers who can use the tool in their day-to-day work.

What are “Sustain.AI’s” quality features?

  • Time saving: The KI.NRW demonstrator allows for quick and efficient handling of sustainability reports and analysis of the required CSR criteria. With the help of AI language models, the text passages relevant to the respective criteria are filtered out. In this way, auditors can focus on the sections that are most relevant to the respective criterion.
  • Overview and context: With the integrated PDF viewer, the extracted text elements can be displayed in the report at any time. This allows users to immediately grasp the context of the passage.  
  • Integrated feedback and adaptability: Users can evaluate the system’s suggestions using the built-in feedback system. This helps us to further train the AI model and continuously improve its quality. It is also capable of learning new criteria.
“Sustainability is increasingly becoming the focus of public attention. With the AI-supported tool Sustain.AI, it is possible to efficiently analyze and browse sustainability reports.”
Maren Pielka
Fraunhofer Institute for Intelligent Analysis and Information Systems IAIS

Which AI technology is used in the KI.NRW demonstrator?

Language models

Modern language models are neural networks which are trained to predict a word by the context in which it appears. In this way, they learn an accurate representation for all words in the text and is capable of semantically comparing and classifying text passages.

Image recognition

PDF is a widely used unstructured file format. This means that although headings and tables are visually recognizable to readers, there is no internal structure. To be able to work with PDFs automatically, intelligent image processing algorithms are required to extract this structure from the images and correctly classify objects such as tables and paragraphs.

What does the AI demonstrator show?

The AI-supported suggestion system is an intelligent, intuitive search engine. You can upload your own documents as well as view existing reports. The reports can be searched and analyzed using a stored checklist from the “Global Reporting Initiative”, a widely used reporting framework for sustainability reporting.

Request a non-binding consultation with our experts now!

Curious? Click here to go straight to the demonstrator!

Where can I find more information? with entries around data analysis and forecasting

More AI providers, applications, and AI products “made in NRW” with the same AI focus can be found using the filter and search function of the AI.Map, which currently contains more than 1000 entries.

Study “Modern language technologies”

Find out where we encounter modern language technologies in everyday life and at work and what economic opportunities they offer.

Fraunhofer IAIS: Media Engineering

You can learn more about the topics of “Cognitive Business Optimization”, “Smart Coding and Learning”, and AI-based industrial image processing on the website of the IAIS Media Engineering department.

Contact the team of developers

Maren Pielka

Data Scientist and Team Leader Cognitive Text Analytics,
Business Area Cognitive Business Optimization,
Media Engineering Department

Fraunhofer IAIS
Schloss Birlinghoven
53757 Sankt Augustin

Phone +49 2241 14-2871

Send email

Lars Patrick Hillebrand

PhD Student / Research Assistant in Machine Learning,
Media Engineering Department

Fraunhofer IAIS
Schloss Birlinghoven
53757 Sankt Augustin

Phone +49 2241 14-1920

Send email

Chatbot with knowledge graph
“Covid Q&A”

The chatbot demonstrates the functionality of an online dialog system that uses the strengths of a knowledge graph.

Where does the AI application offer the greatest benefit?

A chatbot supported by AI offers a wide range of possible applications and assists companies in any area of communication with employees or customers, regardless of the industry. A classic example are customer-oriented service offerings that are made available twenty-four hours a day. Companies use them on websites, in online stores, on support pages, in apps, or via instant messaging systems to simplify navigation on websites, to answer specific customer inquiries, or to structure access to service and customer support.

However, chatbots can be used successfully not only in external but also in internal corporate communications. Application examples can be found in the onboarding of new employees, in HR or administrative processes such as questions about vacation requests or payroll tax settlements, as well as in the support of complex assembly instructions in production.

Chatbots can also be supplemented by the component of acoustic speech recognition as well as acoustic speech synthesis. In this extended form, we speak of voice assistants (voicebot), similar to Siri or Alexa.

What are the quality indicators of such AI applications?

Understanding texts with Natural Language Understanding (NLU)

NLU methods are based on semantic representations of texts. They can understand and map relationships between words. These semantic representations exceed the possibilities of the classical rule-based methods of text mining.

Access information and prepare dialogs with Dialog Management (DM) and Knowledge Graph (KG)

Knowledge Graphs structure data and knowledge, enable semantic linking, and in many cases are the basis for making artificial intelligence applications explainable and providing results that are comprehensible to humans.

Generate texts with Natural Language Generation (NLG)

Text synthesis is the counterpart to text comprehension. Here, text is generated automatically, which can afterwards be transformed into acoustic speech signals.

“Knowledge graphs that integrate different data sources,
form the basis for many AI applications and assistants.”
Prof. Dr. rer. nat Jens Lehmann
Fraunhofer Institute for Intelligent Analysis and Information Systems IAIS

What does the demonstrator of KI.NRW show?

An AI-based dialog system (“question answering system”) shapes the search for information more efficiently and conveniently for the user. The KI.NRW demonstrator shows such an AI-supported chatbot for querying Corona case numbers worldwide. The animations explain step by step how it operates.

Step 1: Understanding natural language

By using AI, speech recognition is oriented to real conversations of humans and the so-called natural language (“natural language understanding NLU”). The models extract information from the text, which they represent internally in a way that allows further processing. This enables the system to understand less common terms, dialects or everyday language. Associations and connections between words are also taken into account appropriately, for example, that the word invoice is related in content to the word payment.

Step 2: Structure data and knowledge

Knowledge graphs structure data and knowledge, enable semantic linking and, in many cases, are the basis for shaping AI applications explainable. A knowledge graph is able to combine a wide variety of information sources into a dynamic knowledge base. In the case of the KI.NRW demonstrator, Corona case numbers from Johns Hopkins University and the Robert-Koch-Institut are included. They are made accessible with the status of the previous day via a knowledge graph.

Step 3: Generate answer

Finally, an answer is generated that matches the asked question. Thus, this process is the matching counterpart to the first step, the understanding of language. The structured data is now converted into text and output. The output as an acoustic signal can also follow here (for example with the so-called voicebots).

The AI chatbot in action

Try out the AI chatbot for yourself: As an example, the chatbot was set up on data that depicts the worldwide case numbers surrounding the Corona pandemic. Anyone who asks a corresponding question in English via the chat window will receive an immediate answer.

Test questions for the knowledge query can be the following:

  • “Are there new cases in Mexico?”
  • “How many cases were there in total in Germany until 25th October 2020?”
  • “How many new cases were found in Argentina on 10th November 2020?”
  • “Which country had the highest number of cases on 8th November 2020?”

Where is more information to be found?

AI provider from NRW

Our AI map illustrates who offers AI methods related to “knowledge and inference” in their portfolio.

AI methods around knowledge and inference

Many “good practice examples” as inspiration for the possible applications of this AI.

AI products “made in NRW”

Filter our AI map by “knowledge and inference”.

Contact us

Roman Teucher

Research Engineer

Fraunhofer IAIS
Zwickauer Str. 46
01069 Dresden

Phone +49 351 85477961

E-Mail senden

To the top