SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2508.00443 · ICCV · 2025

SDMatte: Grafting Diffusion Models for Interactive Matting

Bo Li, Hao Zhang, Wei Dong, Peng-Tao Jiang, Jinwei Chen, Yu Liang, Longfei Huang, Lunde Chen, Wanyu Liu

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 11 functions out of this paper's own repositories and ran 0 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
vivoCameraResearch/SDMatte canonical 0 of 11
FunctionStatusWhere it lives
add_aux_conv_in Not yet run vivoCameraResearch/SDMatte/utils/utils.py
code served (permissive licence) · get_code("589373b8cfdd39d7")
custom_get_attention_scores Not yet run vivoCameraResearch/SDMatte/utils/replace.py
code served (permissive licence) · get_code("44239863af5276ee")
custom_prepare_attention_mask Not yet run vivoCameraResearch/SDMatte/utils/replace.py
code served (permissive licence) · get_code("07f2d45f03a56f84")
dgauss Not yet run vivoCameraResearch/SDMatte/utils/evaluate.py
code served (permissive licence) · get_code("c6daa9009470dd68")
gauss Not yet run vivoCameraResearch/SDMatte/utils/evaluate.py
code served (permissive licence) · get_code("b4085a9b496ef62d")
gauss_kernel Not yet run vivoCameraResearch/SDMatte/criterion.py
code served (permissive licence) · get_code("b4991ebb886e5289")
gaussgradient Not yet run vivoCameraResearch/SDMatte/utils/evaluate.py
code served (permissive licence) · get_code("38441a7c36e020f6")
get_unknown_tensor_from_pred Not yet run vivoCameraResearch/SDMatte/utils/utils.py
code served (permissive licence) · get_code("763a4435e84627a8")
laplacian_loss Not yet run vivoCameraResearch/SDMatte/criterion.py
code served (permissive licence) · get_code("a3fadbcac00e9133")
laplacian_pyramid Not yet run vivoCameraResearch/SDMatte/criterion.py
code served (permissive licence) · get_code("8a0b6d481e748e9d")
replace_unet_conv_in Not yet run vivoCameraResearch/SDMatte/utils/utils.py
code served (permissive licence) · get_code("91e5d642193c6016")

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. Interactive image matting results of our SDMatte with box prompts. SDMatte leverages strong diffusion priors, ensuring robust generalization. Meanwhile, it transforms the text-driven image generation capability of Stable Diffusion into a visual prompt-driven interactive capability, enabling precise alpha matte prediction based on simple user-provided visual prompts (points, boxes, masks).

For agents

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

get_harvested_code_for_paper("2508.00443")
get_code_for_paper("2508.00443")
have("2508.00443")

Connect an agent — have() is free.