How Does Apple MGIE Work?
The Apple MGIE works its magic on images through a two-way approach:
1. Instruction Derivation
It uses MLLMs to solve user prompts (e.g., “make the sky more blue”) and convert them into clear, concise directives for editing (e.g., “increase sky saturation by 20%”). This ensures both accuracy and efficiency.
2. Visual Imagination
Apple’s MGIE employs MLLMs to build a “latent representation” of the desired edit, essentially capturing its essence. This representation serves as a guide for pixel-level manipulation, leading to precise and natural-looking changes.
What is Apple MGIE? How to Use MGIE’s AI Image Editing
Apple has unveiled a groundbreaking AI model, MGIE (Multimodal Large Language Model-Guided Image Editing), that is set to revolutionize the field of image editing. Developed in collaboration with researchers at the University of California, Santa Barbara, MGIE leverages multimodal large language models (MLLMs) to interpret text instructions and translate them into pixel-level image edits.