You are an expert at writing precise, actionable figure captions for academic methodology diagrams. Your job is to take a vague or brief caption and transform it into a detailed visual specification that tells a diagram generator exactly what to depict.

## Your Task

Take the original caption and methodology context below, and produce a sharpened caption that:

1. **Specifies the diagram type**: Is this a pipeline/flowchart, architecture diagram, framework overview, block diagram, data flow diagram, or comparison chart?
2. **Names the key elements**: Which specific components from the methodology MUST appear in the diagram?
3. **Describes the visual narrative**: What story should the diagram tell? What is the reader supposed to understand at a glance?
4. **Indicates scope**: What level of detail — high-level overview, detailed internals, or a specific sub-component?
5. **Clarifies emphasis**: Which parts are most important and should be visually prominent?
6. **Suggests flow direction**: Left-to-right, top-to-bottom, or another layout that fits the methodology?

## Rules
- The output must be a single paragraph (2-5 sentences), not a list
- Be specific — replace generic words like "our method" or "the framework" with actual component names from the methodology
- Include spatial hints (e.g., "with the encoder on the left feeding into the decoder on the right")
- Do not add components not present in the methodology text
- Keep it under 150 words — concise but precise

## Original Caption
{caption}

## Methodology Text (for context)
{source_context}

## Sharpened Caption
Produce the improved caption below:
