AI can generate images given text descriptions

In News by Xu CuiLeave a Comment

AI never fails to surprise me. Last week, OpenAI announced a new AI called “DALL·E” which can generate images based on a short text description. For example,

Input: “an armchair in the shape of an avocado

Then DALL E generated a few versions:

The quality stuns me. DALL E clearly understands the meaning of the text and is creative at the same time. Another example:

Input: “a male mannequin dressed in an orange and black flannel shirt and black jeans

My cousin is a clothes designer and I am sure DALL E will help her design new clothes.

What about unrealistic ideas? Say,

Input: “a snail made of harp

My mind blows away! If I did not know this was created by AI, I would say this is done by a creative artist!

For us who works in the life science industry, I think DALL E will help us too. For example, when we write some content (e.g. a blog post or social media article), we usually need a nice looking image which matches the content. Often we search in Google or stock image websites (such as https://unsplash.com or https://stock.adobe.com), but most of the times the images are not exactly what we look for. With DALL E, we will have a personal creative designer who does exactly what we ask for!

As of today, OpenAI have not made DALL E to the public yet. I really look forward to it.

You should definitely check out more examples in OpenAI’s page: https://openai.com/blog/dall-e/

If you find the article useful, you may consider to subscribe:

Leave a Comment