SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2308.13853 · ICCV · 2023

Beyond One-to-One: Rethinking the Referring Image Segmentation

Wenqi Shao, Ping Luo, Zhenguo Li, Enze Xie, Jungong Han, Yutao Hu, Qixiong Wang

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 13 functions out of this paper's own repositories and ran 9 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
toggle1995/RIS-DMMI canonical 9 of 13
FunctionStatusWhere it lives
window_partition Ran toggle1995/RIS-DMMI/lib/backbone.py
pointer only (licence: NONE) · get_code("144d10b49baeb8a6")
find_pruneable_heads_and_indices Ran toggle1995/RIS-DMMI/bert/modeling_utils.py
pointer only (licence: NONE) · get_code("b239247b87d4f3dc")
gelu Ran toggle1995/RIS-DMMI/bert/modeling.py
pointer only (licence: NONE) · get_code("d08f324f950148de")
get_activation Ran toggle1995/RIS-DMMI/bert/modeling.py
pointer only (licence: NONE) · get_code("5842ee67106321d6")
mish Ran toggle1995/RIS-DMMI/bert/modeling_bert.py
pointer only (licence: NONE) · get_code("22b41155e451683b")
pad_if_smaller Ran toggle1995/RIS-DMMI/transforms.py
pointer only (licence: NONE) · get_code("452475fb6da97402")
prune_conv1d_layer Ran toggle1995/RIS-DMMI/bert/modeling_utils.py
pointer only (licence: NONE) · get_code("b3ee64494a3b6d38")
prune_linear_layer Ran toggle1995/RIS-DMMI/bert/modeling_utils.py
pointer only (licence: NONE) · get_code("a37615738b3ddcd3")
window_reverse Ran toggle1995/RIS-DMMI/lib/backbone.py
pointer only (licence: NONE) · get_code("61bf152e6a42a184")
dmmi_res Not yet run toggle1995/RIS-DMMI/lib/segmentation.py
pointer only (licence: NONE) · get_code("7a437a829041ee13")
dmmi_swin Not yet run toggle1995/RIS-DMMI/lib/segmentation.py
pointer only (licence: NONE) · get_code("6d20e15ac43f4f86")
get_assignment_map_from_checkpoint Not yet run toggle1995/RIS-DMMI/bert/modeling.py
pointer only (licence: NONE) · get_code("50958618b65e514e")
load_tf_weights_in_bert Not yet run toggle1995/RIS-DMMI/bert/modeling_bert.py
pointer only (licence: NONE) · get_code("fb9398de2966d6de")

Repositories linked to this paper

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

Abstract

Referring image segmentation aims to segment the target object referred by a natural language expression. However, previous methods rely on the strong assumption that one sentence must describe one target in the image, which is often not the case in real-world applications. As a result, such methods fail when the expressions refer to either no objects or multiple objects. In this paper, we address this issue from two perspectives. First, we propose a Dual Multi-Modal Interaction (DMMI) Network, which contains two decoder branches and enables information flow in two directions. In the text-to-image decoder, text embedding is utilized to query the visual feature and localize the corresponding target. Meanwhile, the image-to-text decoder is implemented to reconstruct the erased entity-phrase conditioned on the visual feature. In this way, visual features are encouraged to contain the critical semantic information about target entity, which supports the accurate segmentation in the text-to-image decoder in turn. Secondly, we collect a new challenging but realistic dataset called Ref-ZOM, which includes image-text pairs under different settings. Extensive experiments demonstrate our method achieves state-of-the-art performance on different datasets, and the Ref-ZOM-trained model performs well on various types of text inputs. Codes and datasets are available at https://github.com/toggle1995/RIS-DMMI.

For agents

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

get_harvested_code_for_paper("2308.13853")
get_code_for_paper("2308.13853")
have("2308.13853")

Connect an agent — have() is free.