Klaus Schrenker, Business Development Manager
Semiconductor manufacturing involves over a thousand processing stages executed within a specialized and intricate production facility. Couple this with the various production requirements to develop specific wafer and chip varieties, and the lack of a uniform semiconductor manufacturing process becomes apparent. This is driving the industry to seek adaptable solutions to support swift value creation regardless of the disparate nature of procedures. Recently, machine vision has emerged as the go-to technology for consolidating diverse processing tasks, helping semiconductor manufacturers boost their operational efficiencies.
MVTec Software GmbH is at the forefront of providing transformative machine vision software for the semiconductor space. A leading international manufacturer of machine vision software, MVTec supports the demanding areas of image processing across semiconductor production, surface inspection, quality control, and surveillance applications. MVTec serves customers with cutting-edge machine vision software as a crucial link in the value chain.
"We help semiconductor manufacturers achieve new milestones in cost reduction and yield improvement by advancing manufacturing processes and production scenarios using machine vision software," says Klaus Schrenker, business development manager, MVTec Software GmbH.
Through MVTec's machine vision software, users can process images and conduct product inspections in milliseconds. Unlike manual inspection and testing processes plagued by delays due to limited availability of personnel and time-consuming procedures, MVTec's machine vision automation facilitates 24/7 accessibility while showcasing more reliable detection capabilities than the human eye. From a safety standpoint, the software enables users to eliminate t h e risk of performing tasks involving hazardous substances and to seamlessly transition their product inspection processes into clean room settings. Above all, MVTec equips semiconductor manufacturers with a tool that delivers 100 percent objective, accurate, reliable, and repeatable results.
At the heart of these capabilities is MVTec's Munich, Germany-based competence center, where it designs and develops its machine vision software suite, which includes MVTec HALCON, MVTec MERLIC, and Deep Learning Tool. HALCON is the comprehensive standard software for machine vision with a dedicated integrated development environment (HDevelop) tailored toward developing machine vision solutions.
MVTec MERLIC is a user-friendly all-in-one software product for quickly creating machine vision applications without programming. Additionally, the MVTec Deep Learning Tool facilitates data labeling without any programming knowledge providing seamless integration into HALCON and MERLIC. This allows clients to perform deep-learning-based object detection, classification, semantic segmentation, anomaly detection, and Deep OCR.
MVTec's machine vision software features cutting-edge capabilities like a broad collection of techniques for 3D vision and deep learning, besides its remarkable speed and GPU acceleration. HALCON provides the most extensive machine vision library due to its more than 2,100 robust operators. HALCON’s flexible architecture facilitates rapid development and therefore a fast time-to-market of any kind of machine vision application.As customers deploy HALCON in tens of thousands of advanced applications worldwide, it is continuously further developed and refined. To best serve its global customer base, MVTec has subsidiaries and partners all around the world to provide fast and customized services and support, for example in case of issues or production changes.
We help semiconductor manufacturers achieve new milestones in cost reduction and yield improvement by advancing manufacturing processes and production scenarios using machine vision software
As an example, MVTec's machine vision software enabled a European machine manufacturer specialized in backend processing to boost their image processing and defect detection capabilities. The customer sought an inspection application to effectively assess dies for cracks and fractures before the actual bonding. Defects are to be detected at an early stage, so the corresponding dies can be sorted out in advance. With MVTec's HALCON, rule-based and deep learning methods were combined to best solve this challenge. Rule-based methods help to define the evaluation region for pre-processing. The regions of interest are then used to train the neural networks. One challenge in many semiconductor production process steps is the lack of defect images due to the scarcity of defective parts while the variety of potential defects can be large. Using MVTec's Anomaly Detection, only “good images” are required to train the deep learning model which makes it possible to robustly detect any anomalies that do not match the standard. Later and over time, the “bad images” can be used to further analyze and classify the defect types and continuously improve production processes and product quality.
Offering ground-breaking machine vision products and services that promote responsible and efficient resource management, MVTec is making a real difference in the semiconductor market. The company continues to support the machine vision software market, significantly contributing to long-term economic success by drawing on its 25 years of expertise and laser focus on flexibility and quality.