Which will be useful to extract text / information better from ppts containing images?

Question

Which will be useful to extract text / information better from ppts containing images?

Kaushal Chawale 0

I have a ppt and the ppt includes tables and images. In some slides there are multiple images in one slide. These images together form a meaningful data. What should I use Vision Studio or Document Intelligence to extract data from such pages maintaining the relation between the images.

For EG: We have a life cycle of plants. Each stage of the plant is represented by multiple single images. These images have some captions. The images are connected to each other in cyclic arrows. What can I do so that I can extract the information on that slide in text format.

Here,

Input: The plant life cycle.

Output: Explanation of the plant life cycle and that too in a sequential order.

Regards

santoshkc 15,355 Reputation points Microsoft External Staff Moderator

2025-04-11T14:09:54.94+00:00

Hi @Kaushal Chawale,

If your PowerPoint slides have multiple images that work together to tell a story, like the life cycle of a plant with captions and arrows connecting each stage, neither Azure Vision Studio nor Document Intelligence on their own will fully get you the result you're looking for.

Vision Studio can help pull out the text from the images using OCR, and Document Intelligence is good at extracting structured info from forms or tables. But in your case, you need more than that, you want to understand the meaning behind how those images are connected and turn that into a clear explanation in the right order.

The best way to approach this would be to first extract all the text and captions from the slides using OCR (with Vision Studio). Then, you can feed that text along with some context about how the images are connected into a language model like Azure OpenAI’s GPT-4o, etc. That model can then put everything together and generate a step-by-step explanation.

So basically, use Vision Studio to get the raw text, and then use Azure OpenAI model to understand and explain it like a human would.

I hope you understand! Thank you.
santoshkc 15,355 Reputation points Microsoft External Staff Moderator

2025-04-14T10:47:11.3466667+00:00

Hi @Kaushal Chawale,

Following up to see if the given response was helpful. And, if you have any further query do let us know.

Thank you.
santoshkc 15,355 Reputation points Microsoft External Staff Moderator

2025-04-15T12:22:23.3133333+00:00

Hi @Kaushal Chawale ,

We haven’t heard from you on the last response and was just checking back to see if the given response was helpful. Thank you.

Your answer

santoshkc 15,355 Reputation points Microsoft External Staff Moderator

2025-04-11T14:09:54.94+00:00

Hi @Kaushal Chawale,

If your PowerPoint slides have multiple images that work together to tell a story, like the life cycle of a plant with captions and arrows connecting each stage, neither Azure Vision Studio nor Document Intelligence on their own will fully get you the result you're looking for.

Vision Studio can help pull out the text from the images using OCR, and Document Intelligence is good at extracting structured info from forms or tables. But in your case, you need more than that, you want to understand the meaning behind how those images are connected and turn that into a clear explanation in the right order.

The best way to approach this would be to first extract all the text and captions from the slides using OCR (with Vision Studio). Then, you can feed that text along with some context about how the images are connected into a language model like Azure OpenAI’s GPT-4o, etc. That model can then put everything together and generate a step-by-step explanation.

So basically, use Vision Studio to get the raw text, and then use Azure OpenAI model to understand and explain it like a human would.

I hope you understand! Thank you.
santoshkc 15,355 Reputation points Microsoft External Staff Moderator

2025-04-14T10:47:11.3466667+00:00

Hi @Kaushal Chawale,

Following up to see if the given response was helpful. And, if you have any further query do let us know.

Thank you.
santoshkc 15,355 Reputation points Microsoft External Staff Moderator

2025-04-15T12:22:23.3133333+00:00

Hi @Kaushal Chawale ,

We haven’t heard from you on the last response and was just checking back to see if the given response was helpful. Thank you.

Share via

Which will be useful to extract text / information better from ppts containing images?

Your answer