Langdes: A New Approach for Improving the Performance of Prompt-based Image Editing in Interior Design Setting

The topic of instruction-based image editing has gotten a lot of attention in recent years with a lot of research conducted due to its immense potential in various applications such as removing unwanted details present in existing images or improving them. However, one of the main problems in addressing this problem is acquiring a dataset for model training. Several methods and variations were proposed, but all of them rely on already-existent data. We propose a method to address this problem by creating a context-specific dataset for interior design with no previously available information by leveraging the knowledge of large language models (LLM). Furthermore, we test and prove the efficiency of the generated dataset on InstructPix2Pix which starts to compute better results for the interior-design setting after the fine-tuning. Moreover, we propose an alternative solution for enhancing the localization of the edit region through cross-attention map regularization based on a text-based segmentation mask.