Papers › SUM: Saliency Unification through Mamba for Visual Attention Modeling

SUM: Saliency Unification through Mamba for Visual Attention Modeling

25 Jun 2024arXiv:2406.17815archive 2025-07-28

Alireza Hosseini, Amirhossein Kazerouni, Saeed Akhavan, Michael Brudno, Babak Taati

Visual attention modeling, important for interpreting and prioritizing visual stimuli, plays a significant role in applications such as marketing, multimedia, and robotics. Traditional saliency prediction models, especially those based on Convolutional Neural Networks (CNNs) or Transformers, achieve notable success by leveraging large-scale annotated datasets. However, the current state-of-the-art (SOTA) models that use Transformers are computationally expensive. Additionally, separate models are often required for each image type, lacking a unified approach. In this paper, we propose Saliency Unification through Mamba (SUM), a novel approach that integrates the efficient long-range dependency modeling of Mamba with U-Net to provide a unified model for diverse image types. Using a novel Conditional Visual State Space (C-VSS) block, SUM dynamically adapts to various image types, including natural scenes, web pages, and commercial imagery, ensuring universal applicability across different data types. Our comprehensive evaluations across five benchmarks demonstrate that SUM seamlessly adapts to different visual characteristics and consistently outperforms existing models. These results position SUM as a versatile and powerful tool for advancing visual attention modeling, offering a robust solution universally applicable across different types of visual content.

PaperPDFCodeCode Syntology ran

In Syntology View this paper on Syntology: its repositories, every harvested function with whether it ran, its licence and the call to fetch it.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="2406.17815")

Code

Syntology Ran 14 of 14 code samples harvested from 1 repository linked to this paper; 0 have no recorded run. Of those that ran: 14 ran with no contract checked.

By repository: official repository: 14 samples from 1 repository, 14 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Arhosseini77/SUM officialmentioned in papermentioned on GitHubpytorchMIT report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

14 samples harvested; 14 ran; 0 honoured the contract we drafted; 0 have no recorded run. Read from Syntology's graph 2026-09-25; that is when this build read the record, not when the samples ran.

14ran

Licence: 0 of the 14 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from Arhosseini77/SUM. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples from a repository marked official are listed first. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of four reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

flops_selective_scan_ref Arhosseini77/SUM/net/models/vmamba.py official repository ran MIT (permissive) · c00dcf8428ee981b · report
get_logger Arhosseini77/SUM/net/utils.py official repository ran MIT (permissive) · 23a1f34627822bb3 · report
get_optimizer Arhosseini77/SUM/net/utils.py official repository ran MIT (permissive) · ae1a2dd9d62e7fad · report
get_scheduler Arhosseini77/SUM/net/utils.py official repository ran MIT (permissive) · dea42821258184c2 · report
load_and_preprocess_image Arhosseini77/SUM/net/inference_utils.py official repository ran MIT (permissive) · a214ec961739db31 · report
loss_CC Arhosseini77/SUM/utils/loss_function.py official repository ran fingerprinted MIT (permissive) · 774d792a6525842a · report
loss_KLdiv Arhosseini77/SUM/utils/loss_function.py official repository ran fingerprinted MIT (permissive) · e9183c75893d49f0 · report
loss_similarity Arhosseini77/SUM/utils/loss_function.py official repository ran fingerprinted MIT (permissive) · 9596d79dd2070fff · report
mean_std Arhosseini77/SUM/optuna_network.py official repository ran fingerprinted MIT (permissive) · f0c787b1471d4421 · report
modulate Arhosseini77/SUM/net/models/vmamba.py official repository ran fingerprinted MIT (permissive) · 95949c1af8f08e9d · report
modulate_scale Arhosseini77/SUM/net/models/vmamba.py official repository ran MIT (permissive) · e26191359b0c82c3 · report
postprocess_img Arhosseini77/SUM/utils/data_process_uni.py official repository ran MIT (permissive) · ebfa76e7f72447ce · report
predict_saliency_map Arhosseini77/SUM/net/inference_utils.py official repository ran MIT (permissive) · 40e257a513360d15 · report
preprocess_img Arhosseini77/SUM/utils/data_process_uni.py official repository ran MIT (permissive) · f3af41b9e983a72c · report

Tasks

MambaMarketingSaliency DetectionSaliency Prediction

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Saliency Detection CAT2000 SUM AUC 0.888 #1 of 2 Archive leaderboard report
Saliency Detection CAT2000 SUM NSS 2.423 #1 of 2 Archive leaderboard report
Saliency Prediction CAT2000 SUM KL 0.27 #1 of 1 Archive leaderboard report
Saliency Prediction MIT300 SUM AUC-Judd 0.913 #1 of 2 Archive leaderboard report
Saliency Prediction MIT300 SUM CC 0.768 #1 of 2 Archive leaderboard report
Saliency Prediction MIT300 SUM KLD 0.563 #1 of 2 Archive leaderboard report
Saliency Prediction MIT300 SUM NSS 2.839 #1 of 2 Archive leaderboard report
Saliency Prediction MIT300 SUM SIM 0.63 #1 of 2 Archive leaderboard report
Saliency Prediction SALECI SUM KL 0.473 #1 of 5 Archive leaderboard report
Saliency Prediction SALICON SUM AUC 0.876 #1 of 5 Archive leaderboard report
Saliency Prediction SALICON SUM CC 0.909 #1 of 5 Archive leaderboard report
Saliency Prediction SALICON SUM KLD 0.192 #1 of 5 Archive leaderboard report
Saliency Prediction SALICON SUM NSS 1.981 #1 of 5 Archive leaderboard report
Saliency Prediction SALICON SUM SIM 0.804 #1 of 5 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

AttentionConcatenated Skip ConnectionConvolutionDense ConnectionsLayer NormalizationLinear LayerMambaMax PoolingMulti-Head AttentionReLUResidual ConnectionSoftmaxStochastic DepthSwin TransformerU-Net

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections