Fujitsu has developed an innovative AI technology that significantly improves manufacturing quality control and defect detection by automatically analyzing and diagnosing Non-Destructive Testing (NDT) ultrasonic scan data in just minutes, helping to pinpoint potential defects more rapidly and efficiently than existing processes. Fujitsu Laboratories of Europe’s solution uses a new AI framework that combines image and signal processing techniques with deep learning technology to solve complex manufacturing quality problems.
The deep learning component of this new AI framework leverages the super-human ability of deep neural networks to process image data to detect relevant patterns, based on a unique set of technologies developed by Fujitsu Laboratories of Europe. This involves converting real-world data analysis challenges into an image analysis format, automating and accelerating the detection of relevant patterns in NDT ultrasound scan data, which may be indicative of manufacturing defects. Specialist manual inspection can therefore be rapidly targeted to potential defects, translating into an 80 percent reduction in the product area requiring an expert technician’s attention. As a result, quality control is considerably improved and potential bottlenecks in the production process are removed, with the potential to scale up production and make significant efficiency improvements. Additionally, Fujitsu’s solution has the capability to keep learning after deployment, enabling continuous performance improvement and an enhanced return on investment.
Dr Adel Rouz, Executive Vice President of Fujitsu Laboratories of Europe, explains the significance of this innovative technology approach: “We developed a generic machine learning engine for pattern detection, using a process that translates any raw data analysis problem into one involving image pattern recognition. Working with manufacturers, we can rapidly tune the solution to a specific application, thanks to its ability to learn from just a few training examples. This significantly minimizes the amount of annotated data needed from a manufacturer’s domain specialists, accelerating the entire set-up process. At Fujitsu Laboratories of Europe, we are focusing on a co-creation strategy to solve real-world manufacturing problems in record time, using the knowledge base of Fujitsu’s extensive manufacturing expertise combined with our state of the art AI innovations.“
Fujitsu Laboratories of Europe’s technology has already been successfully deployed in a variety of applications, including time-series sensor data, energy consumption, stock price analysis and smart manufacturing. In one application, it was applied to improve the retrieval of 3D CAD models from massive databases, helping to accelerate product design and enhance QC. In another example, the technology was applied to a social innovation application, detecting driver behavior via a discrete wrist-worn acceleration sensor. Potentially dangerous behaviors, such as eating and drinking or programming the GPS navigator while driving, were accurately classified, using a novel method for converting accelerometer time-series data into image representations, which were then fed to a deep neural network.