Stable Diffusion Approaches for Image Generation from Text in Low-Resource Languages
Abstract
Diffusion-based generative models have recently emerged as powerful frameworks for computational vision tasks, particularly for generating images from textual descriptions. Among these, Stable Diffusion has become a widely adopted approach for producing high-quality images from textual inputs expressed in natural language. Despite this progress, most current text-to-image systems mainly target high-resource languages like English. This focus leaves low-resource languages relatively unexplored. Consequently, challenges persist in terms of limited language representation, difficulties in aligning text and images, a lack of data, and issues with cultural and semantic accuracy in generated images. This survey provides a structured review of text-to-image generation methods for low-resource languages, concentrating on techniques based on Stable Diffusion and related diffusion architectures. It reviews how prior studies adapt these models for low-resource situations using methods like multilingual and cross-lingual language encoders, translation-based workflows, parameter-efficient fine-tuning, and knowledge distillation. We also summarize commonly used datasets and evaluation metrics found in the literature, along with their limitations in low-resource environments. By organizing and summarizing previous work, this survey aims to highlight current trends and outline unresolved challenges in creating inclusive text-to-image generation systems for low-resource languages.
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.