Thursday, 4 December 2025
25.2 C
Singapore
27.9 C
Thailand
22.6 C
Indonesia
27.9 C
Philippines

Apple unveils revolutionary AI image editing model

Discover Apple's groundbreaking AI image editing model, revolutionizing editing with simple text commands and superior performance.

Apple has stepped into AI image editing by releasing a groundbreaking open-source multimodal AI model. Developed in collaboration with researchers from the University of California, Santa Barbara, the MLLM-Guided Image Editing (MGIE) model promises to revolutionize how images are edited.

Simplifying image editing with natural language commands

The MGIE model enables users to edit images using simple text commands, akin to giving instructions to a Photoshop expert. Unlike traditional methods, which often struggle to comprehend brief instructions, MGIE leverages multimodal large language models (MLLMs) to simultaneously understand natural language prompts and image data. This innovative approach eliminates the need for detailed descriptions, resulting in more accurate and intuitive editing.

According to the research paper accompanying the release, existing AI image editing tools frequently fall short when interpreting human instructions. However, MGIE’s ability to grasp the context of expressive instructions sets it apart. For instance, when presented with an image of a pepperoni pizza and instructed to “make this more healthy,” MGIE intelligently infers that “this” refers to the pizza and interprets “more healthy” as adding vegetables. The result? A pepperoni pizza adorned with nutritious green toppings, transforming it into a healthier option.

Superior performance demonstrated

In comparative tests, MGIE outperformed other models by accurately implementing complex editing commands. For example, when instructed to “add lightning and make the water reflect the lightning” to an image of a forested shoreline, MGIE seamlessly incorporated the lightning and its reflection on the water’s surface. In contrast, competing models failed to capture the intricate details of the command, highlighting MGIE’s exceptional capabilities.

Apple’s foray into AI image editing marks a significant step forward in the field, demonstrating the company’s commitment to advancing AI technologies. While the release of MGIE is a notable development, industry observers speculate that this may be just the beginning for Apple in the realm of AI.

Hot this week

Honor launches Magic8 Pro in Singapore with new MagicBook Art 14 and Watch Fit

Honor launches the Magic8 Pro in Singapore with upgraded imaging, AI features and companion devices including the MagicBook Art 14 and Watch Fit.

Red Hat expands AWS collaboration to enhance AI inference performance

Red Hat expands its AWS collaboration to support large-scale generative AI with improved performance and lower costs.

Kayou debuts at Singapore Comic Con 2025 with focus on Southeast Asia expansion

Kayou marks its debut at Singapore Comic Con 2025 and outlines plans to expand its retail network and fan community efforts across Southeast Asia.

Honor showcases early low-light camera performance of the Magic 8 Pro

Honor offers an early look at the Magic 8 Pro’s upgraded low-light camera performance during brief testing at the Singapore Oceanarium.

AI browsers vulnerable to covert hacks using simple URL fragments, experts warn

Experts warn AI browsers can be hacked with hidden URL fragments, posing risks invisible to traditional security measures.

HPE expands AI-native networking portfolio and outlines vision for self-driving IT operations

HPE expands its AI-native networking portfolio with new AIOps features, hardware, and hybrid cloud tools designed for self-driving IT operations.

Kyndryl and Microsoft report rising sustainability commitment among Singapore businesses

Most Singapore businesses are expanding sustainability efforts but face challenges with data quality and limited AI adoption.

UnionBank adopts Amazon Quick Suite to accelerate data-driven decision making

UnionBank deploys Amazon Quick Suite to expand access to data analytics and speed up decision making across its organisation.

Kargo Technologies outlines plan for 40,000-vehicle EV shift by 2035

Kargo Technologies sets a 2035 target to deploy 40,000 electric vehicles and build an AI-driven Electrified Silk Road across Asia.

Related Articles

Popular Categories