CLIPScore
PulseAugur coverage of CLIPScore — every cluster mentioning CLIPScore across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
New framework fools CLIP-based image quality metrics, defense proposed
Researchers have developed FoCLIP, a framework designed to manipulate and detect adversarial examples for CLIP-based image quality assessment. This method intentionally misaligns features in the image-text space to arti…
-
New method balances emotion and semantics in text-to-image generation
Researchers have developed a new method for controlling emotions in text-to-image generation that aims to improve affective alignment without altering the core semantic content of the image. This approach uses a flow-ma…
-
New SVG-Score framework improves evaluation of text-to-SVG generation models
Researchers have introduced SVG-Score, a new evaluation framework designed to better assess the quality of text-to-SVG generation models. Existing metrics like CLIPScore, originally developed for natural images, are not…
-
New SVG-Score framework improves evaluation for text-to-SVG generation
Researchers have introduced SVG-Score, a new evaluation framework specifically designed for text-to-SVG generation, addressing the limitations of existing metrics like CLIPScore which are ill-suited for vector graphics.…
-
New benchmark evaluates AI's ability to illustrate poetry
Researchers have developed TangPoetryBench, a new benchmark designed to evaluate text-to-image generation models specifically for their ability to illustrate poetry. Existing metrics fall short in capturing nuances like…
-
New CASCADE framework enhances backdoor detection in multimodal learning
Researchers have developed a new framework called CASCADE to detect backdoor attacks in multimodal contrastive learning (MCL). Existing methods often rely on the CLIPScore metric, but this approach has limitations due t…
-
New SAGE method improves safety alignment in text-to-image models
A new research paper published on arXiv introduces StructureAware Geometric Regularization (SAGE), a novel method for improving the safety alignment of text-to-image diffusion models. Current alignment techniques often …
-
New EPIG method enhances emotional expression in AI image generation
Researchers have developed EPIG, a novel method to enhance emotional expressiveness in text-to-image generation. EPIG enriches prompts with emotion-related components based on psychological representations like valence-…
-
Image-text metrics fail semantic invariance tests, researchers find
Researchers have identified significant semantic invariances in popular image-to-text evaluation metrics. These metrics, including CLIPScore and others, show sensitivity to benign spatial edits and phrasing changes, lea…
-
New pipeline generates synthetic industrial defects to augment scarce real-world data
Researchers have developed SynSur, an end-to-end pipeline for generating synthetic industrial surface defects to address the scarcity of labeled data in defect detection. The pipeline combines vision-language models, Lo…