{"id":23222,"date":"2025-06-30T10:40:00","date_gmt":"2025-06-30T10:40:00","guid":{"rendered":"https:\/\/www.compass-horizon.eu\/?p=23222"},"modified":"2025-11-20T15:21:26","modified_gmt":"2025-11-20T15:21:26","slug":"advancing-ai-driven-defect-detection-in-carbon-fiber-materials","status":"publish","type":"post","link":"https:\/\/www.compass-horizon.eu\/?p=23222","title":{"rendered":"Advancing AI-Driven Defect Detection in Carbon Fiber Materials | Technical paper"},"content":{"rendered":"\n

Segmentation Supervised Multi-Head Defect Classification of Virgin and Recycled Carbon Fiber Material in Photometric Stereo Images<\/h3>\n\n\n\n
A newly developed segmentation-supervised multi-head neural network is setting new benchmarks in defect classification for virgin and recycled carbon fiber materials.

With the increasing adoption of neural networks in digital image-based inspection, researchers face challenges due to carbon fiber’s non-rigid nature and high reflectivity, which complicate robust quality assessment.

To address these difficulties, this approach integrates a multi-head, multi-task neural architecture<\/strong> with optimized convolutional layers, enabling shared embeddings for image classification and segmentation<\/strong> in photometric stereo imaging. The framework enhances defect classification by utilizing class activation maps (CAMs) and segmentation outputs<\/strong> as attention mechanisms. Performance evaluations conducted on a publicly available dataset show that this model surpasses standard CNNs and transformer-based methods, achieving over 30% higher accuracy<\/strong>, exceeding previous benchmarks by 4.4%<\/strong>.

Key Innovations:<\/strong>
– Modular optical sensor<\/strong>: An extended camera sensor with photometric stereo capabilities measure fiber orientation parameters.
– Improved classification accuracy<\/strong>: CAM-based attention mechanisms enhance defect classification performance.
– Multi-head architecture<\/strong>: A lightweight architecture processes image classification and segmentation concurrently.
– Benchmark performance<\/strong>: Extensive evaluations on publicly available datasets validate the model\u2019s superior defect detection.

An extended photometric stereo camera setup<\/strong> was developed, capturing multi-modal images that undergo stitching and pre-processing for further analysis. The segmentation architecture applies transposed convolutional layers<\/strong> to extract detailed segmentation features, which are fine-tuned with CAM layers. These segmentations are binarized and used as masks<\/strong>, refining convolution feature extraction for classification tasks.

Conclusions and Future Directions<\/strong>:
This research introduces a comprehensive framework for defect classification<\/strong>, integrating modular optical sensing, attention-based classification refinements, and a lightweight multi-head architecture. The proposed model improves accuracy while reducing reliance on large datasets, a challenge in industrial applications.

Next Steps<\/strong>: The study acknowledges limitations in dataset availability, restricting model generalization across diverse industrial scenarios. Future efforts will focus on image generation using few-shot learning<\/strong>, creating synthetic defect samples to strengthen model scalability.

For access to the dataset and additional technical details, visit Zenodo<\/a>. \u00a0<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Note: <\/strong>The Zenodo links<\/a>, this is a dataset that has been created during the MC4 project<\/a>. All the experiments conducted for the paper used this dataset which is the first and only available dataset in the world for this category. The paper<\/a> (presented at Sampe Europe 2025<\/a>) shares the acknowledgements of both the MC4<\/a> and COMPASS<\/a> projects.<\/p>\n\n\n\n

<\/p>\n","protected":false},"excerpt":{"rendered":"

Segmentation Supervised Multi-Head Defect Classification of Virgin and Recycled Carbon Fiber Material in Photometric Stereo Images A newly developed segmentation-supervised multi-head neural network is setting new benchmarks in defect classification for virgin and recycled carbon fiber materials. With the increasing adoption of neural networks in digital image-based inspection, researchers face challenges due to carbon fiber’s […]<\/p>\n","protected":false},"author":2,"featured_media":23226,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1,33],"tags":[],"class_list":["post-23222","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-news","category-insights"],"_links":{"self":[{"href":"https:\/\/www.compass-horizon.eu\/index.php?rest_route=\/wp\/v2\/posts\/23222","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.compass-horizon.eu\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.compass-horizon.eu\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.compass-horizon.eu\/index.php?rest_route=\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.compass-horizon.eu\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=23222"}],"version-history":[{"count":9,"href":"https:\/\/www.compass-horizon.eu\/index.php?rest_route=\/wp\/v2\/posts\/23222\/revisions"}],"predecessor-version":[{"id":23273,"href":"https:\/\/www.compass-horizon.eu\/index.php?rest_route=\/wp\/v2\/posts\/23222\/revisions\/23273"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.compass-horizon.eu\/index.php?rest_route=\/wp\/v2\/media\/23226"}],"wp:attachment":[{"href":"https:\/\/www.compass-horizon.eu\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=23222"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.compass-horizon.eu\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=23222"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.compass-horizon.eu\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=23222"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}