SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2305.11173 · ICCV · 2023

Going Denser with Open-Vocabulary Part Segmentation

Ping Luo, Saining Xie, Peize Sun, Shoufa Chen, Fanyi Xiao, Chenchen Zhu, Zhicheng Yan

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 5 functions out of this paper's own repositories and ran 2 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
facebookresearch/vlpart canonical 2 of 5
FunctionStatusWhere it lives
window_partition Ran facebookresearch/vlpart/vlpart/modeling/backbone/swintransformer.py
pointer only (licence: MIT) · get_code("f9fd6241d935f07b")
window_reverse Ran facebookresearch/vlpart/vlpart/modeling/backbone/swintransformer.py
pointer only (licence: MIT) · get_code("fb32094c6dbece71")
center_of_mass Not yet run facebookresearch/vlpart/vlpart/modeling/meta_arch/parser_net_mixer.py
pointer only (licence: MIT) · get_code("1f97465a9cdcc617")
chunk_cosine_sim Not yet run facebookresearch/vlpart/vlpart/modeling/meta_arch/parser_net_mixer.py
pointer only (licence: MIT) · get_code("a11a4946af0a8f49")
create_timm_resnet Not yet run facebookresearch/vlpart/vlpart/modeling/backbone/timm.py
pointer only (licence: MIT) · get_code("3634f2208a0255d7")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Figure 1. Examples of open-vocabulary part segmentation. Beyond open-vocabulary object detection, we propose that the detector should be able to predict both objects and their parts. This open-world fine-grained recognition ability is in demand for an intelligent vision system but is only realized in a limited number of categories [9, 29, 67] up to now. In this paper, we move forward to going denser with open-vocabulary part segmentation: Left figure shows segmenting dog and its parts in different granularities. Right figure demonstrates more visualization results.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2305.11173")
get_code_for_paper("2305.11173")
have("2305.11173")

Connect an agent — have() is free.