Presentation Slide
Presentation slide research focuses on automating the creation and understanding of slides, addressing challenges in accessibility and efficiency. Current efforts leverage large language models (LLMs) and vision-language models (VLMs), often integrated with graph neural networks (GNNs) to handle the non-linear structure of presentations and multimodal content (text, images, charts). This work aims to improve accessibility for visually impaired individuals, enhance teaching effectiveness through interactive lecture systems, and streamline the slide creation process for researchers and educators, ultimately increasing productivity and knowledge dissemination.
Papers
Leveraging weak complementary labels to improve semantic segmentation of hepatocellular carcinoma and cholangiocarcinoma in H&E-stained slides
Miriam Hägele, Johannes Eschrich, Lukas Ruff, Maximilian Alber, Simon Schallenberg, Adrien Guillot, Christoph Roderburg, Frank Tacke, Frederick Klauschen
From slides (through tiles) to pixels: an explainability framework for weakly supervised models in pre-clinical pathology
Marco Bertolini, Van-Khoa Le, Jake Pencharz, Andreas Poehlmann, Djork-Arné Clevert, Santiago Villalba, Floriane Montanari