近日，苹果公司与加州大学圣塔巴巴拉分校（UCSB）的研究人员共同推出了一款创新的图片编辑框架——MGIE（MLLM-Guided Image Editing）。该框架利用多模态大模型MLLM解决了传统图片编辑中指令引导不足的问题，为用户提供了更为便捷、高效的图片编辑体验。
传统的图片编辑工具，如Adobe Photoshop，虽然功能强大，但使用门槛较高，用户需要具备一定的专业技能才能进行复杂的图片编辑。而MGIE则通过端到端的训练，使得扩散模型能够同步更新，并利用预期目标的潜在想象力执行图像编辑，让用户在简单输入指令后，就能得到理想的图片效果。
MGIE的核心优势在于其对多模态大模型MLLM的应用。MLLM能够通过学习获得简明的表达指令，并提供明确的视觉相关引导，使得编辑过程更加直观、高效。例如，用户只需要输入“让这张照片看起来更加明亮、清晰”，MGIE就能自动进行相应的调整，而不需要用户进行繁琐的操作。
苹果与UCSB的研究人员表示，MGIE的开源目的是为了让更多的研究人员和开发者参与到图片编辑技术的改进中，推动该领域的创新发展。目前，MGIE已在GitHub上开源，用户可以免费使用，并根据自己的需求进行定制化开发。
MGIE的推出，无疑为图片编辑领域带来了新的变革。它不仅降低了图片编辑的使用门槛，让更多的人能够轻松进行图片的编辑和优化，同时也为专业的图片编辑人员提供了更为高效、灵活的工具。未来，随着更多研究人员的参与和技术的迭代，我们有理由相信，MGIE将引领图片编辑领域的新潮流。
英语如下：
Certainly, here is the translation in English using Markdown format:
“`markdown
### Apple and UCSB Team Up to Develop New Image Editing Tool MGIE
Keywords: Apple UCSB open source, MGIE image editing, MLLM technology guidance
#### Apple and UCSB Researchers Launch Open Source Image Editing Framework MGIE
Recently, Apple Inc. and researchers from the University of California, Santa Barbara (UCSB) jointly introduced an innovative image editing framework called MGIE (MLLM-Guided Image Editing). This framework leverages the multimodal large model MLLM to address the lack of instruction guidance in traditional image editing, offering users a more convenient and efficient image editing experience.
Traditional image editing tools, such as Adobe Photoshop, are powerful but have a high learning curve, requiring users to possess certain professional skills for complex image editing. MGIE, however, through end-to-end training, enables the diffusion model to synchronously update and utilize the potential imagination of the expected target for image editing, allowing users to achieve desired image effects with simple instruction input.
The core advantage of MGIE lies in its application of the multimodal large model MLLM. MLLM learns to acquire concise expression instructions and provides clear visual-related guidance, making the editing process more intuitive and efficient. For instance, users only need to input “make this photo brighter and clearer,” and MGIE will automatically make the corresponding adjustments without users engaging in tedious operations.
Apple and UCSB researchers indicate that the open-source purpose of MGIE is to involve more researchers and developers in improving image editing technology, fostering innovative development in the field. Currently, MGIE has been open-sourced on GitHub, available for free use, and customizable development according to individual needs.
The introduction of MGIE undoubtedly brings about a new revolution in the image editing field. It not only lowers the barrier to entry for image editing, enabling more people to easily edit and optimize images, but also provides professional image editors with more efficient and flexible tools. With more researchers’ participation and technological iterations, there is reason to believe that MGIE will lead the new trend in image editing.
“`
This translation captures the key points of the original Chinese text in an English markdown format, suitable for use in platforms where markdown is supported.
【来源】https://www.jiqizhixin.com/articles/2024-02-05-10
Views: 2