Dynamic Perceiver for Efficient Visual Recognition
Author:
Affiliation:
1. Tsinghua University,BNRist,Department of Automation
2. Tsinghua University,BNRist,Department of Computer Science and Technology
3. China Mobile Research Institute
Funder
National Natural Science Foundation of China
Tsinghua University
Publisher
IEEE
Link
http://xplorestaging.ieee.org/ielx7/10376473/10376477/10377175.pdf?arnumber=10377175
Reference81 articles.
1. Beit: Bert pre-training of image transformers;Bao
2. Batch-shaping for learning conditional channel gated networks;Bejnordi
3. Adaptive neural networks for efficient inference;Bolukbasi
4. CF-ViT: A General Coarse-to-Fine Method for Vision Transformer
Cited by 7 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献
1. Comparison of Machine Learning Approaches for Robust and Timely Detection of PPE in Construction Sites;Robotics;2024-02-16
2. Framework-agnostic Semantically-aware Global Reasoning for Segmentation;2024 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV);2024-01-03
3. Fine-Grained Recognition With Learnable Semantic Data Augmentation;IEEE Transactions on Image Processing;2024
4. Multi-Resolution Model Compression for Deep Neural Networks: A Variational Bayesian Approach;IEEE Transactions on Signal Processing;2024
5. Deep Incubation: Training Large Models by Divide-and-Conquering;2023 IEEE/CVF International Conference on Computer Vision (ICCV);2023-10-01
1.学者识别学者识别
2.学术分析学术分析
3.人才评估人才评估
"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370
www.globalauthorid.com
TOP
Copyright © 2019-2024 北京同舟云网络信息技术有限公司 京公网安备11010802033243号 京ICP备18003416号-3