IDMT-ISAAC - Develop your own AI to improve quality assurance

When industrial manufacturing processes are audible, the quality of these processes can be determined by their sound characteristics. The experienced machine operator hears whether a manufacturing process is successful and is therefore an important part of quality control. However, increasing automation is creating new challenges for quality management in industry. Inspection tasks are becoming more complex and trained personnel are less in demand. How can manufacturing companies react to the advancing automation of their quality control without overloading employees? And how much AI expertise is needed to apply automated quality analysis and interpret the results correctly? Many industrial plants face exactly these questions.

IDMT-ISAAC - Industrial Sound Analysis for Automated Quality Control

With IDMT-ISAAC, Fraunhofer IDMT would like to offer a tool for the automated analysis of industrial sounds using AI. The software framework contains pre-trained AI models and methods based on the analysis of acoustic sensor data. Even users without AI expert knowledge can use this audio analysis tool to achieve fast and reliable results for their quality assurance and to make prompter decisions regarding quality assessment.

IDMT-ISAAC is currently accessible through a browser and the acoustic sensor data chosen for analysis can be imported, simply and securely. Users can try out the pre-built AI models included in the framework, edit them via an intuitive interface and apply them to their own acoustic inspection tasks. Depending on the application, the well-structured dashboard provides an overview of your live data as well as incoming sensor data and helps with further analyses.

Software is looking for applicants

The software tool is currently under development. If you are interested in using the described functions of IDMT-ISAAC, please contact us and together we will find an innovative solution for your application in quality assurance.

• In-line testing of production processes
• Predictive maintenance of machines and plants
• End-of-line testing of industrial products

• Import of acoustic time signals in various formats, e.g. as wave-file
• Configuration of specific extraction and model parameters
• Automatic pre-selection of suitable audio analysis tools and AI models
• Hosted by Fraunhofer IDMT with secure data location
• Server software for internal operation
• Evaluation and export of models and analysis results

• Automatic classification and analysis of industrial sounds
• Pre-trained AI models for audio analysis
• Visualization of analysis results with easy export option
• Easy integration of the developed AI models into the own quality control