Page under construction
AI Vision is the feature to describe an image and convert the image to a prompt. It’s done by using one of the popular IA models such ChatGpt or Gemini. You can select the model you prefer, based on the performance or the cost.
The artificial intelligence AI Vision feature is available for now in Light Image Editor application. Which is a bundled with the installation package of Light Image Resizer. It’s a local app for Windows only. You will need an API key , either for Google Gemini , or ChatGpt from OpenAI. It means, you will pay as you go , no need for montly subscription.
Primo, Get API key of vision models
Get API Key for Google AI studio ( Gemini )
Get API key from OpenAI ( ChatGpt )
Per ora, we recommand to use Gemini, cheaper and works fine for describing images.
Disclaimer : Be careful, even if you are running an local application on windows, the app resize and send the image to the cloud, if you don’t want to share your pictures or sensitive data ( especially nsfw content ) don’t use the cloud models.
Once you have the key ( and credits ), you can enter the key in the Light Image Resizer AI vision option available in this menu.
The LLM vision models works fine in english, francese, tedesco, and popular languages, which is not the case of some other popular vision models, Ollama could be supported later if you find it interesting, and the big advantage means the data will be full private and local, and you won’t be charged for any requests and tokens. See the vision models supported by Ollama
Why describing your pictures ?
Finding the right words for your pictures is easy with Light Image Resizer’s AI Vision feature In Light Image Editor. It analyzes your images at a technical level to automatically generate descriptions. The application prepare your image ( by converting to JPEG and resizing to a size like 1024×1024, which is adjustable) for efficient AI processing. This pre-analysis step helps ensure faster processing and can make the feature more cost-effective. Let the AI do the describing so you don’t have to!
What kind of tasks you can complete ?
Annotate your images to find them by keywords.
Geoguessing the picture when you don’t remember where you take them.
Find inspiration when you want to publish a new instagram or tiktok post. It will add some nice icons to make your marketing better.
It’s all depends on your prompt, see a few samples of prompts and how it impacts the result.