| Towards a Deep Multimodal Similarity Learning for Text and Image Embeddings Fusion |
 |
  |
A precise image classifier based on CNN and a feature extractor, structured as a complete online pipeline with preprocessing, postprocessing, evaluation, testing, training, and visualization modules. Includes a “zero-shot” text classifier using a transformer. |
| Open Source IoT Middleware AI-Based Robot Control for High-Mix, Low-Volume Production Sites |
 |
 |
Development of an open-source IoT data integration platform, applicable to sectors such as healthcare, supply chain, and retail. It enables real-time tracking of objects and people via technologies like Bluetooth Low Energy or RFID and processes data in a standardized way. With its Pareto middleware, reelyActive now provides actionable data and neutral expertise to support organizations’ sustainable digital transformation. |
