AI PHOTO EDITING MADE EASY: APPLE’S MGIE MODEL TAKES TEXT PROMPTS TO IMAGES

Apple has released a new AI model that lets users describe in plain language what they want to change in a photo without ever touching photo editing software.

It’s a MGIE model, which stands for MLLM-Guided Image Editing. It enables image manipulation including cropping, resizing, flipping, and applying filters via text prompts. This tool allows users to describe desired alterations directly to an image when using the Apple MGIE model.

This model is good at managing both simple and complex image alterations, allowing for the modification of particular elements within a photograph, including reshaping objects or boosting their brightness.

Platforms such as DALL-E 3 from OpenAI are capable of conducting basic photo editing tasks on images they generate, using textual commands.

Adobe Firefly is an AI-based editing model. It facilitates the creation of generative fill, enabling the addition of AI-generated backgrounds to photographs.

Example used in Apple’s MGIE paper.

The MGIE documentation showcases the model’s adaptability with examples, such as transforming a pepperoni pizza picture into a healthier version by adding vegetable toppings through the prompt “make it more healthy.”

Additionally, enhancing the brightness of a poorly lit Sahara tigers photo is possible by commanding the model to “add more contrast to simulate more light.

Apple made MGIE available to download via GitHub. as well as releasing a demo on Hugging Face Spaces.