Filters








4 Hits in 4.1 sec

ReCLIP: A Strong Zero-Shot Baseline for Referring Expression Comprehension [article]

Sanjay Subramanian, William Merrill, Trevor Darrell, Matt Gardner, Sameer Singh, Anna Rohrbach
2022 arXiv   pre-print
We present ReCLIP, a simple but strong zero-shot baseline that repurposes CLIP, a state-of-the-art large-scale model, for ReC.  ...  Training a referring expression comprehension (ReC) model for a new visual domain requires collecting referring expressions, and potentially corresponding bounding boxes, for images in the domain.  ...  We thank the Berkeley NLP group, Medhini Narasimhan, and the anonymous reviewers for helpful comments. We thank Michael Schmitz for help with AI2 infrastructure.  ... 
arXiv:2204.05991v2 fatcat:t7td2epzvbe6zmblwthlhtoiri

ReCLIP: A Strong Zero-Shot Baseline for Referring Expression Comprehension

Sanjay Subramanian, William Merrill, Trevor Darrell, Matt Gardner, Sameer Singh, Anna Rohrbach
2022 Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)   unpublished
We present ReCLIP, a simple but strong zero-shot baseline that repurposes CLIP, a state-of-the-art large-scale model, for ReC.  ...  Training a referring expression comprehension (ReC) model for a new visual domain requires collecting referring expressions, and potentially corresponding bounding boxes, for images in the domain.  ...  We thank the Berkeley NLP group and Medhini Narasimhan for helpful comments.  ... 
doi:10.18653/v1/2022.acl-long.357 fatcat:dbhovjdqsbd5vpgcrv6pjytvly

Sim-To-Real Transfer of Visual Grounding for Human-Aided Ambiguity Resolution [article]

Georgios Tziafas, Hamidreza Kasaei
2022 arXiv   pre-print
Experimental results show that the decoupled nature of our framework allows for easy integration with domain adaptation approaches for Sim-To-Real visual recognition, offering a data-efficient, robust,  ...  In this work, we seek to tackle these limitations by introducing a fully decoupled modular framework for compositional visual grounding of entities, attributes, and spatial relations.  ...  Foundation models still offer a strong baseline, especially considering their zero-shot capabilities, although still lacking when compared with domain-specific architectures for a given task. " Grasp the  ... 
arXiv:2205.12089v2 fatcat:bnepnsepwrfxxdtqf4mpgvaube

SANP - Supplementum 1

2010 Swiss Archives of Neurology and Psychiatry  
The protocol consisted of a T1weighted turbo spin echo sequence for anatomical reference ( fig. 1) , two gradient echo sequences for 2-point Dixon imaging [2], a multi-contrast TSE sequence for quantification  ...  A detailed pattern of individual muscle involvement may help on the understanding of the disease and serve as a baseline for longitudinal studies.  ...  Conclusion: DWI is a strong predictor for clinical outcome in BAO treated by IAT.  ... 
doi:10.4414/sanp.2010.02212 fatcat:2ashaeol5zfyvicqlmvzpyuq74