j-min/DSG
Davidsonian Scene Graph (DSG) for Text-to-Image Evaluation (ICLR 2024)
This project offers a robust method to assess the quality of images generated by text-to-image AI models. You input a text prompt and the image produced by an AI, and it outputs a detailed score indicating how well the image matches the text's specific details and nuances. This tool is designed for AI researchers, developers of generative AI models, or anyone needing to rigorously evaluate the faithfulness and consistency of AI-generated visuals against their textual descriptions.
105 stars. No commits in the last 6 months.
Use this if you need an automatic, fine-grained, and reliable way to evaluate whether AI-generated images accurately reflect the precise details and relationships specified in text prompts.
Not ideal if you are looking for subjective aesthetic evaluations or a quick, high-level quality check that doesn't require deep semantic analysis.
Stars
105
Forks
7
Language
Jupyter Notebook
License
Apache-2.0
Category
Last pushed
Dec 09, 2024
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/j-min/DSG"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
zai-org/CogVideo
text and image to video generation: CogVideoX (2024) and CogVideo (ICLR 2023)
zhaorw02/DeepMesh
[ICCV 2025] Official code of DeepMesh: Auto-Regressive Artist-mesh Creation with Reinforcement Learning
YangLing0818/RPG-DiffusionMaster
[ICML 2024] Mastering Text-to-Image Diffusion: Recaptioning, Planning, and Generating with...
thu-nics/FrameFusion
[ICCV'25] The official code of paper "Combining Similarity and Importance for Video Token...
Yushi-Hu/tifa
TIFA: Accurate and Interpretable Text-to-Image Faithfulness Evaluation with Question Answering