Open AI model to generate image as output based on input image

Jyothsna 0 Reputation points
2024-07-15T13:47:10.86+00:00

I have recently started using azure open AI service and i have a scenario where i have to send a image to open ai and get the image as output with proper tags. For example we have a image of a park and we ask open AI to mark benches in the park and return that image as a response. Currently gpt4o model is analyzing image and providing text output whereas i need as a image as output.

Azure OpenAI Service
Azure OpenAI Service
An Azure service that provides access to OpenAI’s GPT-3 models with enterprise capabilities.
2,878 questions
0 comments

2 answers

Sort by: Most helpful
Most helpful Newest Oldest
  1. Azar 22,860 Reputation points MVP
    2024-07-15T17:27:39.6566667+00:00

    Hi there Jyothsna

    Thanks for using QandA platform.

    I guess the Azure OpenAI Service currently focuses on text generation and analysis, and while GPT-4 can analyze images and provide text-based responses, it does not natively support generating or modifying images directly as output. but, you can do this by combining Azure services with a computer vision model capable of image processing.

    https://video2.skills-academy.com/en-us/azure/ai-services/computer-vision/

    https://video2.skills-academy.com/en-us/azure/ai-services/computer-vision/quickstarts-sdk/client-library?pivots=programming-language-python&tabs=windows%2Cvisual-studio

    https://video2.skills-academy.com/en-us/azure/storage/blobs/storage-quickstart-blobs-python?tabs=managed-identity%2Croles-azure-portal%2Csign-in-azure-cli&pivots=blob-storage-quickstart-scratch

    If this helps kindly accept the answer thanks much.

    0 comments
  2. Jyothsna 0 Reputation points
    2024-07-18T06:17:28.75+00:00

    Open AI:

    i see that chat GPT can generate image outputs for input image. Azure open ai will also have this feature in the future?

    Azure Vision Service:

    Azure's Computer Vision service detect the objects in images but it just returns coordinates of objects and not the image.

Your answer

Answers can be marked as Accepted Answers by the question author, which helps users to know the answer solved the author's problem.