Apple has introduced its image editing AI model, although it's still in its early stages of development. Dubbed MGIE, which stands for multimodal large language models-guided image editing, the model is currently accessible on GitHub, accompanied by a project paper in PDF format that provides instructions for use.
Utilizing text instructions, MGIE facilitates alterations and enhancements to images. It can execute basic editing tasks such as adjusting contrast or brightness and modifying white balance. Furthermore, it demonstrates capability in interpreting more intricate prompts, such as enhancing the healthiness of a pizza by incorporating vegetable toppings.
Additionally, the model offers functionalities like cropping and resizing images, autofilling image borders, modifying attributes of subjects such as hair, eyes, and clothing, and removing objects from images.
While MGIE represents an initial stride in Apple's venture into generative AI, it is unlikely to be integrated into current Apple devices. Nevertheless, its introduction serves as a glimpse into future developments in this domain.
Source |
0 Comments