Papers › SALI: Short-term Alignment and Long-term Interaction Network for Colonoscopy Video...

SALI: Short-term Alignment and Long-term Interaction Network for Colonoscopy Video Polyp Segmentation

19 Jun 2024arXiv:2406.13532archive 2025-07-28

Qiang Hu, Zhenyu Yi, Ying Zhou, Fang Peng, Mei Liu, Qiang Li, Zhiwei Wang

Colonoscopy videos provide richer information in polyp segmentation for rectal cancer diagnosis. However, the endoscope's fast moving and close-up observing make the current methods suffer from large spatial incoherence and continuous low-quality frames, and thus yield limited segmentation accuracy. In this context, we focus on robust video polyp segmentation by enhancing the adjacent feature consistency and rebuilding the reliable polyp representation. To achieve this goal, we in this paper propose SALI network, a hybrid of Short-term Alignment Module (SAM) and Long-term Interaction Module (LIM). The SAM learns spatial-aligned features of adjacent frames via deformable convolution and further harmonizes them to capture more stable short-term polyp representation. In case of low-quality frames, the LIM stores the historical polyp representations as a long-term memory bank, and explores the retrospective relations to interactively rebuild more reliable polyp features for the current segmentation. Combing SAM and LIM, the SALI network of video segmentation shows a great robustness to the spatial variations and low-visual cues. Benchmark on the large-scale SUNSEG verifies the superiority of SALI over the current state-of-the-arts by improving Dice by 2.1%, 2.5%, 4.1% and 1.9%, for the four test sub-sets, respectively. Codes are at https://github.com/Scatteredrain/SALI.

PaperPDFConference PDFCode

Code

Scatteredrain/SALI officialmentioned in paperpytorch 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

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

SegmentationVideo Polyp SegmentationVideo SegmentationVideo Semantic Segmentation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Video Polyp Segmentation SUN-SEG-Easy (Unseen) SALI Dice 0.825 #3 of 18 Archive leaderboard report
Video Polyp Segmentation SUN-SEG-Easy (Unseen) SALI S measure 0.870 #3 of 18 Archive leaderboard report
Video Polyp Segmentation SUN-SEG-Easy (Unseen) SALI Sensitivity 0.811 #3 of 18 Archive leaderboard report
Video Polyp Segmentation SUN-SEG-Easy (Unseen) SALI mean E-measure 0.920 #3 of 18 Archive leaderboard report
Video Polyp Segmentation SUN-SEG-Easy (Unseen) SALI mean F-measure 0.831 #3 of 18 Archive leaderboard report
Video Polyp Segmentation SUN-SEG-Easy (Unseen) SALI weighted F-measure 0.794 #3 of 18 Archive leaderboard report
Video Polyp Segmentation SUN-SEG-Hard (Unseen) SALI Dice 0.822 #3 of 18 Archive leaderboard report
Video Polyp Segmentation SUN-SEG-Hard (Unseen) SALI S-Measure 0.874 #3 of 18 Archive leaderboard report
Video Polyp Segmentation SUN-SEG-Hard (Unseen) SALI Sensitivity 0.830 #3 of 18 Archive leaderboard report
Video Polyp Segmentation SUN-SEG-Hard (Unseen) SALI mean E-measure 0.920 #3 of 18 Archive leaderboard report
Video Polyp Segmentation SUN-SEG-Hard (Unseen) SALI mean F-measure 0.822 #3 of 18 Archive leaderboard report
Video Polyp Segmentation SUN-SEG-Hard (Unseen) SALI weighted F-measure 0.790 #3 of 18 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

Absolute Position EncodingsAdamAttentionBPEConvolutionDeformable ConvolutionDense ConnectionsDropoutFocusLabel SmoothingLayer NormalizationLinear LayerMulti-Head AttentionPosition-Wise Feed-Forward LayerResidual ConnectionSAMSoftmaxTransformer

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