Text and Image Automatic Generation Applied to Digital Signage
Abstract
This thesis tackles the challenge of automating content generation for digital signage by leveraging artificial intelligence techniques. It introduces contentAI, a modular and scalable platform designed to integrate seamlessly with existing digital signage infrastructures, enabling the automatic creation of context-aware textual and visual content. The proposed solution strategically combines state-of-the-art generative models, in- cluding large language models such as LLaMA 2 for text and diffusion-based models like FLUX and Stable Diffusion for image generation. These models were selected for their strong performance in terms of language fluency and visual realism, as well as their suitability for domain-specific fine-tuning. While fine-tuning improved the text model’s performance, the fine-tuned version of Stable Diffusion failed to produce consistently better visuals. Instead, FLUX was chosen as the final image model due to its superior perceptual quality, as demonstrated both quantitatively and through human evaluation. A Human-in-the-Loop mechanism was integrated into the system to continuously collect user feedback via a REST API, enabling iterative improvement of the models. This approach allows the platform to evolve dynamically, aligning its outputs with audience expectations and commercial objectives. A key architectural decision was the separation of text and image generation processes, driven by practical limitations in positioning and styling embedded text.…
Read the paper
Related studies
- Learning from First-Generation Qualitative Approaches in the IS Discipline: An Evolutionary View and Some Implications for Authors and Evaluators (PART 1/2)
- Commentary on the clinical and preclinical dosage limits of interstitially administered magnetic fluids for therapeutic hyperthermia based on current practice and efficacy models
- Prosthesis: New Technological Opportunities and Innovative Biomedical Devices
- The history and future of neural modeling for cochlear implants
- Ryanodine receptors: waking up from refractoriness
- Novel Transgenic Rabbit Model Sheds Light on the Puzzling Role of Matrix Metalloproteinase-12 in Atherosclerosis
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.