SceneNapse (Pronounced, like "Synapse")
A Google VISTA-inspired prompt optimization framework for text-to-image generation using SigLIP2-based quality scoring and DSPy's GEPA optimizer.
Core trick
Scores generated images with SigLIP2 embeddings and feeds the score back into a DSPy GEPA optimizer that rewrites the prompt.
Description
A Google VISTA-inspired prompt optimization framework for text-to-image generation using SigLIP2-based quality scoring and DSPy's GEPA optimizer.
README excerpt
# SceneNapse Studio **Multi-head DSPy pipeline for cinematic scene composition and text-to-image generation with quality scoring and prompt alignment verification.** --- ## Complete Workflow SceneNapse operates in two phases: **Prompt Enhancement** transforms text into structured scene descriptions, then **Image Generation & Validation** produces and verifies images against the prompt. ### Phase 1: Prompt Enhancement <p align="center"> <img src="frontend/public/workflow-infographic.jpeg" width="600" alt="Prompt Enhancement Pipeline"> </p> User prompts (with optional multimodal inputs) are decomposed into a structured **Scene Ontology** with four parallel heads: Objects, Actions, and Cinematography—all grounded in the Elements layer. ### Phase 2: Image Generation & Validation <p align="center"> <img src="frontend/public/workflow-infographic2.jpeg" width="400" alt="Image Generation & Validation Pipeline"> </p> The enriched scene description drives image generation (Nano Banana Pro), with **JoyQuality** scoring candidates and **VLM Guardrails** verifying prompt alignment. --- ## Overview SceneNapse Studio transforms simple text prompts into **structured, composable scene descriptions** using a 3-stage DSPy pipeline. Each scene is decomposed into four independent "heads" (Elements, Objects, Actions, Cinematography) that can be refined individually or regenerated together. This architecture enables: - **Precise control** over visual elements, motion, and cinemato