Apple has recently developed a new artificial intelligence model with the University of California, Santa Barbara. With the help of this model, named MGIE (MLLM-Guided Image Editing), users can easily edit their photos by simply providing plain-language text commands.
The model can perform tasks such as cropping, resizing, rotating, and adding filters to images. First, the model learns how to interpret user commands and then “imagines” what the image would look like after the edit. For instance, if a user asks for a bluer sky, the model increases the brightness in the sky portion of the image. This makes the photo editing process quite simple and straightforward, as the user only needs to mention what they want to change about the image.
“Multimodal large language models (MLLMs) show promising capabilities in cross-modal understanding and visual-aware response generation via LMs. We investigate how MLLMs facilitate edit instructions and present MLLM-Guided Image Editing (MGIE). MGIE learns to derive expressive instructions and provides explicit guidance. The editing model jointly captures this visual imagination and performs manipulation through end-to-end training. We evaluate various aspects of Photoshop-style modification, global photo optimization, and local editing. Extensive experimental results demonstrate that expressive instructions are crucial to instruction-based image editing, and our MGIE can lead to a notable improvement in automatic metrics and human evaluation while maintaining competitive inference efficiency.” states the research paper, which is published as a conference paper at ICLR, 2024 (International Conference on Learning Representations 2024).
Apple has made MGIE available for download on GitHub and also released a web demo on Hugging Face. However, when compared to its competitors, such as Adobe’s Firefly and OpenAI’s ChatGPT, Apple still lags behind in artificial intelligence products.
