[Feature Request]When using the advanced analysis options, I'd like the ability to use the original image as the answer.

## Describe the solution you'd like

When using the advanced analysis options, I'd like the ability to use the original image as the answer.

## Why the solution needed

For multimodal data such as flow charts, displaying the original image as is is easier to understand and use than having users respond in text.

## Additional context

I tried to implement it by modifying the sample source, but it seems difficult to do so because the scope of modifications is large, such as:
bedrock-claude-chat/backend/app/vector_search.py
bedrock-claude-chat/frontend/src/components/RelatedDocumentViewer.tsx
bedrock-claude-chat/backend/app/routes/s3_proxy.py *Added
bedrock-claude-chat/backend/app/main.py
bedrock-claude-chat/backend/app/utils.py
etc.

## Implementation feasibility

Are you willing to collaborate with us to discuss the solution, decide on the approach, and assist with the implementation?

- [ ] Yes, I am able to implement the feature and create a pull request.
- [X] No, I am unable to implement the feature, but I am open to discussing the solution.

Trace the advanced analysis flow across backend/app/vector_search.py, backend/app/routes/s3_proxy.py, backend/app/main.py, backend/app/utils.py, and frontend/src/components/RelatedDocumentViewer.tsx. First determine how the original image could travel through the backend and be rendered by the frontend. Done means users can select the original image as the answer for multimodal content such as flow charts.

Written by the indexing model from the issue text.