@Billy Sun , Thanks for additional clarity.
The Azure Cognitive Search blob indexer can extract text PPT/PPTX and other document formats, listed in this document. The indexer will open the file and extract text, images, and metadata
Azure Cognitive Search can’t return page numbers to you by default, you may try these:
Return hit highlights of the search results and compare the text in the hit highlights with the doc to identify which page(s) in the document the match is on. Also, see Paging results for more info.
Additionally - Extract text and information from images in AI enrichment scenarios
This article covers image processing in more detail and provides guidance for working with images in an AI enrichment pipeline.