Training-Free Location-Aware Text-to-Image Synthesis
This addresses a key limitation in generative models for users needing fine-grained spatial control, though it is incremental as it builds on existing stable diffusion models.
The paper tackles the problem of precise control over object size and position in text-to-image synthesis by proposing a training-free interactive generation paradigm, achieving superior performance in control capacity and image quality compared to state-of-the-art methods.
Current large-scale generative models have impressive efficiency in generating high-quality images based on text prompts. However, they lack the ability to precisely control the size and position of objects in the generated image. In this study, we analyze the generative mechanism of the stable diffusion model and propose a new interactive generation paradigm that allows users to specify the position of generated objects without additional training. Moreover, we propose an object detection-based evaluation metric to assess the control capability of location aware generation task. Our experimental results show that our method outperforms state-of-the-art methods on both control capacity and image quality.