ResearchBenchmarksVisual Text Generation on MuST-Bench EnglishVisual Text Generation on MuST-Bench EnglishFollow71.63OCR AccuracySIGIL31.79842.13952.4862.821Oct 24, 2024Evaluation resultsMethodMethodLinksOCR AccuracyCLIP-I ScoreSIGIL2024.1071.6386.73DS-Fusion2024.1048.0885.96DALL-E32024.1043.682.1AnyText2024.1033.3387.31