Docling’s PPTX parser does not support extracting images embedded inside placeholders or grouped shapes directly currently as there's no built in options like pipeline_options.generate_picture_images for PPTX.

This limitation exists due to the fact Docling relies on on parsing the PPTX XML structure, as well as its PPTX pipeline being simpler than its PDF pipeline. Images inside placeholders or grouped shapes are also nested deeper inside complex XML relationships and are not exposed as standalone picture elements.