Adobe Photoshop is widely acknowledged as a pivotal software, utilized by over 90% of creative professionals globally, as reported by Photutorial. Recently, Alibaba’s Qwen Team has introduced a new open-source AI model, Qwen-Image Edit, which can perform a variety of Photoshop-like editing tasks using text-based instructions. This model builds upon the 20-billion-parameter Qwen-Image foundation model released earlier this month and aims to enhance text rendering capabilities in image editing.
To process an image, users simply upload a starting picture and provide text commands for desired changes. Qwen-Image Edit will then generate a modified image in response. The model is available across several platforms, including Qwen Chat, Hugging Face, and Alibaba Cloud API, facilitating integration into various applications.
This AI model supports both English and Chinese inputs, and it features a dual-encoding mechanism that can modify the semantic meaning of images while also preserving their visual fidelity. This allows for nuanced editing, including both broad changes, like object transformation, as well as specific alterations, such as removing hair strands or adjusting image text.
Qwen-Image Edit is positioned as a versatile tool for different applications, including creative design, advertising, personal photography, and cultural preservation. It is reported to deliver state-of-the-art performance in image editing, although detailed benchmark figures were not included.
Developers can access the model through Alibaba Cloud’s API at a pricing rate of $0.045 per image, with a free quota for initial use. The service aims to lower barriers for visual content creation, which may benefit both professional designers and casual users. As AI models transition from single-purpose functions to integrated editing capabilities, Qwen-Image Edit represents this evolutionary step.
Source: https://venturebeat.com/ai/qwen-image-edit-gives-photoshop-a-run-for-its-money-with-ai-powered-text-to-image-edits-that-work-in-seconds/