Edit Text and Objects Within Images Using Semantic Control
Modify existing images and update text without starting from scratch.
Leverage Qwen-Image-Edit to perform precise image manipulations, specifically for changing text within an image while maintaining visual consistency.
The Scenario
You have a great stock photo or a previously generated AI image, but you need to update the text on a sign or change a specific visual element to fit a new campaign.
Before & after
You would traditionally use the 'Clone Stamp' or 'Content-Aware Fill' tools in Photoshop to remove or change text, then try to match the lighting and texture manually. This takes 15–20 minutes of careful editing.
Utilize Qwen-Image-Edit to describe the specific text or object you want to alter. The model performs precise semantic and appearance editing in about 1–2 minutes without needing manual masking.
The Prompt
[UPLOAD_IMAGE] In this image, please change the text that says '[OLD_TEXT]' to say '[NEW_TEXT]'. Maintain the original font style, lighting, and perspective.
This model uses Qwen2.5-VL for visual semantic control and a VAE Encoder for appearance control, allowing for seamless edits that maintain the original image's integrity.
Source
Blog | Qwen"Qwen-Image-Edit successfully extends Qwen-Image’s unique text rendering capabilities to image editing tasks, enabling precise text editing."