↓Skip to main content

🏢 Brookhaven National Laboratory

CLIPCEIL: Domain Generalization through CLIP via Channel rEfinement and Image-text aLignment
·3674 words·18 mins· loading · loading
AI Generated Multimodal Learning Vision-Language Models 🏢 Brookhaven National Laboratory
CLIPCEIL enhances CLIP’s domain generalization by refining feature channels for domain invariance and aligning image-text embeddings, achieving state-of-the-art performance.