Výsledky vyhledávání

Report

A Simple Image Segmentation Framework via In-Context Examples

Autor: Liu, Yang, Jing, Chenchen, Li, Hengtao, Zhu, Muzhi, Chen, Hao, Wang, Xinlong, Shen, Chunhua

Recently, there have been explorations of generalist segmentation models that can effectively tackle a variety of image segmentation tasks within a unified in-context learning framework. However, these methods still struggle with task ambiguity in in

Externí odkaz: http://arxiv.org/abs/2410.04842

Zobrazit plný text záznamu

Report

Unleashing the Potential of the Diffusion Model in Few-shot Semantic Segmentation

Autor: Zhu, Muzhi, Liu, Yang, Luo, Zekai, Jing, Chenchen, Chen, Hao, Xu, Guangkai, Wang, Xinlong, Shen, Chunhua

The Diffusion Model has not only garnered noteworthy achievements in the realm of image generation but has also demonstrated its potential as an effective pretraining method utilizing unlabeled data. Drawing from the extensive potential unveiled by t

Externí odkaz: http://arxiv.org/abs/2410.02369

Zobrazit plný text záznamu

Report

Generative Active Learning for Long-tailed Instance Segmentation

Autor: Zhu, Muzhi, Fan, Chengxiang, Chen, Hao, Liu, Yang, Mao, Weian, Xu, Xiaogang, Shen, Chunhua

Recently, large-scale language-image generative models have gained widespread attention and many works have utilized generated data from these models to further enhance the performance of perception tasks. However, not all generated data can positive

Externí odkaz: http://arxiv.org/abs/2406.02435

Zobrazit plný text záznamu

Report

DiverGen: Improving Instance Segmentation by Learning Wider Data Distribution with More Diverse Generative Data

Autor: Fan, Chengxiang, Zhu, Muzhi, Chen, Hao, Liu, Yang, Wu, Weijia, Zhang, Huaqi, Shen, Chunhua

Instance segmentation is data-hungry, and as model capacity increases, data scale becomes crucial for improving the accuracy. Most instance segmentation datasets today require costly manual annotation, limiting their data scale. Models trained on suc

Externí odkaz: http://arxiv.org/abs/2405.10185

Zobrazit plný text záznamu

Report

De novo protein design using geometric vector field networks

Autor: Mao, Weian, Zhu, Muzhi, Sun, Zheng, Shen, Shuaike, Wu, Lin Yuanbo, Chen, Hao, Shen, Chunhua

Innovations like protein diffusion have enabled significant progress in de novo protein design, which is a vital topic in life science. These methods typically depend on protein structure encoders to model residue backbone frames, where atoms do not

Externí odkaz: http://arxiv.org/abs/2310.11802

Zobrazit plný text záznamu

Report

SegPrompt: Boosting Open-world Segmentation via Category-level Prompt Learning

Autor: Zhu, Muzhi, Li, Hengtao, Chen, Hao, Fan, Chengxiang, Mao, Weian, Jing, Chenchen, Liu, Yifan, Shen, Chunhua

Current closed-set instance segmentation models rely on pre-defined class labels for each mask during training and evaluation, largely limiting their ability to detect novel objects. Open-world instance segmentation (OWIS) models address this challen

Externí odkaz: http://arxiv.org/abs/2308.06531

Zobrazit plný text záznamu

Report

Matcher: Segment Anything with One Shot Using All-Purpose Feature Matching

Autor: Liu, Yang, Zhu, Muzhi, Li, Hengtao, Chen, Hao, Wang, Xinlong, Shen, Chunhua

Powered by large-scale pre-training, vision foundation models exhibit significant potential in open-world image understanding. However, unlike large language models that excel at directly tackling various language tasks, vision foundation models requ

Externí odkaz: http://arxiv.org/abs/2305.13310

Zobrazit plný text záznamu