{"id":880,"date":"2022-05-01T07:38:00","date_gmt":"2022-05-01T05:38:00","guid":{"rendered":"https:\/\/www.profactor.at\/?post_type=project&#038;p=880"},"modified":"2026-04-04T15:48:53","modified_gmt":"2026-04-04T13:48:53","slug":"flexspect-ai","status":"publish","type":"project","link":"https:\/\/www.profactor.at\/en\/project\/flexspect-ai\/","title":{"rendered":"FlexSpect.AI"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\">FlexSpect.AI addresses a key Industry 4.0 challenge: reliable, automated visual quality inspection (AVQI) in flexible manufacturing. For high-value goods in the packaging, plastics, and printing industries, precise quality control is essential. In practice, however, existing systems are often costly, complex to configure, and heavily dependent on human expertise. This results in long setup times, higher error susceptibility, avoidable scrap, and increased energy consumption.   <\/p>\n\n<p class=\"wp-block-paragraph\">One key reason is that machines do not \u201csee\u201d quality the way humans do. Traditional machine vision and AI models require large amounts of training data and are sensitive to product changes, new designs, or modified materials. At the same time, human inspectors bring subjective judgments, fatigue, and inconsistent decision-making. FlexSpect.AI addresses this gap by developing an AI-enabled system that translates human aesthetic perception into a robust, quantifiable technology for automated quality control\u2014faster, more stable, and more resource-efficient.   <\/p>\n\n<h3 class=\"wp-block-heading\">Project objectives and technical innovation<\/h3>\n\n<p class=\"wp-block-paragraph\">The goal of FlexSpect.AI is to develop a new generation of automated visual quality inspection systems that drastically reduce human effort and can be rapidly adapted to new products and designs.<\/p>\n\n<p class=\"wp-block-paragraph\">Core innovations include:<\/p>\n\n<ul class=\"wp-block-list\">\n<li><strong>Human Aesthetic Perception Module<\/strong>: A module that translates subjective human quality judgments into measurable parameters. This makes \u201cvisual quality\u201d\u2014such as print appearance, colour, and surface finish\u2014consistently quantifiable for AI algorithms. <\/li>\n\n\n\n<li><strong>Robust domain adaptation and transfer-learning concept<\/strong>: Existing data are intelligently reused instead of starting from scratch each time. The approach addresses the key dimensions of transfer learning: changing source domains (e.g., new materials or designs), shifting distance metrics, and hyperparameter optimization. <\/li>\n\n\n\n<li><strong>Active learning to reduce human involvement<\/strong>: The system requests human input only in genuinely critical cases. This minimizes labeling effort and setup time while enabling the model to improve continuously. <\/li>\n<\/ul>\n\n<p class=\"wp-block-paragraph\">Practical applicability will be demonstrated through two proof-of-concept use cases: <strong>high-speed printing of packaging materials<\/strong> and <strong>plastic molding of 3D parts<\/strong>. In addition, best-practice guidelines will be developed to enable the transfer of FlexSpect.AI to other industries. <\/p>\n\n<h3 class=\"wp-block-heading\">AI-based quality inspection: less scrap, greater efficiency<\/h3>\n\n<ul class=\"wp-block-list\">\n<li><strong>For manufacturing companies<\/strong>: FlexSpect.AI significantly reduces the effort required to set up and adapt AVQI systems when switching products. Robust transfer learning shortens downtime, increases process stability, and cuts scrap as well as resource waste. <\/li>\n\n\n\n<li><strong>For quality and energy efficiency<\/strong>: Consistent, objective quality decisions\u2014independent of shift, individual operator, or fatigue\u2014improve product quality and enable more efficient use of materials, energy, and machine capacity.<\/li>\n\n\n\n<li><strong>For Industry 4.0 and AI in manufacturing<\/strong>: The project delivers a transferable reference framework for \u201cQuality AI\u201d that can be integrated across a wide range of industries\u2014from packaging and plastics to other areas of discrete manufacturing.<\/li>\n<\/ul>\n\n<p class=\"wp-block-paragraph\">Coordinated by <strong>PROFACTOR GmbH<\/strong> and carried out together with the <strong>Software Competence Center Hagenberg<\/strong>, <strong>adapa Holding GesmbH<\/strong>, and <strong>Greiner Packaging International GmbH<\/strong>, FlexSpect.AI demonstrates how advanced AI, transfer learning, and active learning can redefine automated visual quality inspection in the smart factories of the future.<\/p>\n","protected":false},"featured_media":686,"template":"","categories":[70,51],"class_list":["post-880","project","type-project","status-publish","has-post-thumbnail","hentry","category-automation-for-flexible-manufacturing","category-industry-5-0"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.profactor.at\/en\/wp-json\/wp\/v2\/project\/880","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.profactor.at\/en\/wp-json\/wp\/v2\/project"}],"about":[{"href":"https:\/\/www.profactor.at\/en\/wp-json\/wp\/v2\/types\/project"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.profactor.at\/en\/wp-json\/wp\/v2\/media\/686"}],"wp:attachment":[{"href":"https:\/\/www.profactor.at\/en\/wp-json\/wp\/v2\/media?parent=880"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.profactor.at\/en\/wp-json\/wp\/v2\/categories?post=880"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}