Projects
Camera-based In-Line Inspection of Machine Parts
In this project, a software was developed that inspects the surface of machine parts. Based on camera images, the software detects and evaluates surface anomalies like scratches or pores. The evaluation result determines whether machine parts are accepted or rejected.
Technologies: Python, PyPylon, OpenCV, PyTorch
Open‑Source Plugins for Metadata Management of the Verfassungsblog
The Verfassungsblog is an academic Open‑Access journal discussing topics regarding constitutional law in Germany. In this project, WordPress plugins were developed to improve the management of article metadata and allow exporting information in multiple formats, e.g., as Marc21-XML.
Technologies: PHP, WordPress, MySQL, XML, XSLT
Development Support for the Open‑Source Software “Kitodo.Production”
Kitodo.Production is a web-based tool to support digitalization work in libraries, museums and digital archives, e.g., importing and managing scans of documents or books, compiling and exporting metadata. In this project, the software was improved regarding both performance and usability.
Technologies: Java, PrimeFaces, MySQL, ElasticSearch
REST Service for Document Classification in Digital Libraries
In this project, the Open‑Source software “slub_docsa” was developed. It allows the comparison of state-of-the-art machine learning methods (e.g., Artificial Neural Networks) for the purpose of automatically annotating documents with detailed class hierarchies often used by libraries (e.g., DDC). A REST service provides the trained models.
Technologies: Python, Scikit-Learn, PyTorch, Transformers
Adaptation and Maintenance of the Search Tool “BLE‑X”
BLE‑X is a web-based research tool that allows to search within the full text of research reports. Search results are visualized in groups of similar documents to support users during the research process. In this project, the Java application was updated according to new requirements.
Technologies: Java, Lucene, JavaScript, jQuery, Docker
Demos
Document Classification Service
Utilize modern machine learning methods to automatically determine suitable classes or topics for a document based on simple information like its title, abstract or full text.
Atlas: Innovative Document Search
Atlas is a novel document search tool that combines the power of AI with classic full-text search. Documents are automatically analyzed for topics, which then can be explored in an interactive virtual map.
We look Forward to Hearing from You
Send a Message
Call by Phone
+49-391-40594560
(Mon ‒ Fri, 9am ‒ 5pm)