You can now edit DALL-E generated images within ChatGPT, announced OpenAI. The image editing feature is available across the web, iOS, and Android. Additionally, ChatGPT now shows style inspirations for DALL-E images within the interface.

For the uninitiated, DALL-E is a text-to-image generation tool from OpenAI powered by generative AI. It is integrated within ChatGPT, but available only to ChatGPT Plus subscribers. Until now, ChatGPT allowed users to generate images based on text prompts. Now, OpenAI enabled the option to edit DALL-E generated images with the ChatGPT platform.

For image editing, OpenAI has introduced a new editor interface that includes ‘using the selection tool’ – which can be accessed by clicking on an image generated by DALL-E. With the selection tool, users can select an area in the image to edit and then describe the changes in chat in the form of a text prompt. Besides, users can provide a prompt with the desired edit in the conversation panel.

OpenAI recommends selecting a large space around the edit area for better results. In the image editing tool interface, there is also an option to undo, redo and clear selection to make changes or restart editing. The interface allows adding, removing and updating parts of an image. After giving the prompts, the image can be saved by selecting the “Save” button on the top-right corner of the editor.

Besides this, there is another option to edit the image by simply giving prompts. To use this method, users need to give desired prompts in the conversational panel on the right side of the DALL-E editor interface.

Editing with DALL-E on the ChatGPT mobile app

ChatGPT mobile app

Similar to the web, users can edit the images on the DALL-E interface on the ChatGPT mobile app – available for Android and iOS. On the smartphone app, select the image generated and it will reveal four options for editing – Edit, Select, Save, and Share. By choosing the ‘select’ option, users can access the selection tool and highlight the parts of the image they want to edit. The slider on the left side of the tool’s interface allows you to change the size of the selection tool. Like the web client, there are undo and redo buttons available at the bottom of the screen to undo and redo selections. After the editing is done, users can click ‘Next’ at the bottom-right of the screen. There, users can give prompts for the desired changes they want in the highlighted areas of the image.




First Published: Apr 04 2024 | 5:32 PM IST

Source link