TerraVis: Towards Evaluation of World-Grounded Visual Consistency in Text-to-Image Generation via MLLM Workflows
cs.CV
Submitted: 2026-10-02
Updated: 2026-10-02
Code: https://github.com/ShyFoo/TerraVis
Terminology
Sources
- HunyuanImage 3.0 Technical Report
- Emerging Properties in Unified Multimodal Pretraining
- Counting Hallucinations in Diffusion Models
- GenAI-Bench: Evaluating and Improving Compositional Text-to-Visual Generation
- Hierarchical Text-Conditional Image Generation with CLIP Latents
- OpenAI GPT-5 System Card
- Gemini: A Family of Highly Capable Multimodal Models
- Gemma 4 Technical Report
- SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Features
- Emu3: Next-Token Prediction is All You Need
- Qwen-Image Technical Report
- Human Preference Score v2: A Solid Benchmark for Evaluating Human Preferences of Text-to-Image Synthesis
- Scaling Autoregressive Models for Content-Rich Text-to-Image Generation
Related papers
- Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
- AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
- Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
- MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
- TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
- A Survey on Efficient Vision-Language-Action Models