Le Générateur d'Objets
In the object generator, the creation of images by Artificial Intelligence is done in 2 stages , in order to structure the request (the "prompt") sent to the AI in a more precise and efficient way:
the choice of the Universe , and the choice of the Subject .
choice of the universe
Choose the aesthetic universe (Post-Impressionism, Bauhaus, Baroque, Gothic, modern, etc.), you will then have the choice between selecting a predefined universe by clicking on a thumbnail, or writing your own universe.
Choice of the subject
The choice of subject takes place in the same way as for the Universe. You have the choice between selecting a proposed subject by clicking on a thumbnail, or freely writing the subject of your choice.
Please note : Our AI understands French, but English often gives more convincing results. The predefined choices (the thumbnails) are automatically sent to the AI in English.
Requests containing prohibited terms (e.g. attempts to generate pornographic images) will be automatically rejected and generate an error.
You will then arrive on a page presenting the universe and the chosen subject , and a confirmation button. It is on this same page that you can select a format . Our artificial intelligence model is in fact able to generate a visual in many formats, from extreme panoramic to vertical.
It is not simply a “cut” of the image, the AI will actually adapt the placement of the objects or figures represented in the image so as to return a harmonious image.
For example, if you plan to order a T-shirt , it is recommended that you choose a " portrait " format but this is not obligatory.
Once the format has been chosen, you can press the “ Imagine ” button before starting to calculate the image.
We decided to prioritize quality over any other variable, which is why generating the image can take a little time: generally around 15 or 20 seconds. This is a deliberate choice on our part and it is the same for all advanced AI today.
At the end of the process, the result is displayed. You then have several possibilities:
- The “ Validate creation ” button will save the image in our system. The visual products (digital image files) will be generated, as well as the objects (the products) which will be illustrated by the image that you have generated by AI: t-shirts, posters, etc. And you will be automatically redirected to the store, on the visual files product page. From this same product page of the store, you will find all the links allowing you to access all versions of the item (sweatshirt, phone case, etc.).
Please note that it may take up to 1-2 minutes for all items in the store to generate, this is due to the time it takes for the server to generate them. If, after clicking on a product link, an error such as "this page does not exist" is displayed, this is due to the fact that the creation of the product has not yet been completed. In this case we suggest waiting a few moments before trying again.
Regenerate the image . This button allows you to generate a new version of the visual, using the same parameters.
You can also press the " Restart " symbol (rotating arrow) to start from scratch.
Our AI model
We use an in-house Artificial Intelligence model, with automatic adaptation of the engine according to the artistic universe, in order to obtain the most aesthetic result possible. AI is still in its infancy, and we are continually updating our models to use the most advanced AI.
It can happen that the AI is "capricious", it cannot be controlled absolutely: it can only be influenced. Therefore, the further the chosen subject is from the universe, the more likely the AI will be unable to connect the two. So, in this case, it can sometimes happen that the subject seems absent from the image.
The AI does not yet know how to produce images with transparent backgrounds. If you need to print a design with a transparent background, instead of a square or rectangular image, do not hesitate to call on our retouching service .
Evolution of the model used by The Object, from v1 to the current version.
Prompt: A painting by Vermeer, hot-air balloon
AI-generated images on The Object are royalty-free for personal or commercial use.