Research map / Video / Video segmentation and tracking
Drop-DTW: Aligning Common Signal Between Sequences While Dropping Outliers
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| jointly drop outliers and align sequences with Drop-DTW instead of DTW-based baselines | instructional-video step localization · CrossTask, COIN, YouCook2 | IoU: 30.8 -> 34.2; 27.7 -> 29.5; 47.5 -> 49.4 | Helped |
| replace percentile drop costs with learned drop costs (ablation) | Drop-DTW step localization · CrossTask | IoU: 34.2 -> 36.9 | Helped |
| add clustering loss to Drop-DTW training (ablation) | Drop-DTW step localization · CrossTask | Recall: 33.8 -> 48.9 | Helped |
| add clustering loss to SmoothDTW training (ablation) | SmoothDTW step localization · CrossTask | Recall: 29.6 -> 43.1 | No clear effect |
| use learned drop cost rather than percentile cost (ablation) | Drop-DTW · CrossTask | Recall: 48.9 -> 49.7 | Helped |
| vary percentile drop-cost setting from p = 0.3 to p = 0.9 (ablation) | Drop-DTW · CrossTask | Recall: 48.9 -> 44.6 | Hurt |
| replace SmoothMin with SoftMin (ablation) | Drop-DTW · CrossTask | Recall: 48.9 -> 46.1 | Hurt |
| replace SmoothMin with hard min (ablation) | Drop-DTW · CrossTask | Recall: 48.9 -> 48.3 | No clear effect |
| optimize dropping and alignment jointly instead of dropping before DTW | step localization inference · CrossTask | IoU: 13.8 -> 30.5 | Helped |
| use Drop-DTW instead of SmoothDTW as the audio-visual sequence matching cost | unsupervised AVE localization · AVE A2V | A2V: 39.8 -> 41.1 | Helped |
| use Drop-DTW instead of SmoothDTW as the audio-visual sequence matching cost | unsupervised AVE localization · AVE V2A | V2A: 33.9 -> 35.8 | Helped |
| use Drop-DTW rather than SmoothDTW for representation learning with contaminated sequences | PennAction · PennAction with interspersed outliers | Kendall's Tau | Helped |
Closest papers by meaning
- Deep Declarative Dynamic Time Warping for End-to-End Learning of Alignment Paths (ICLR 2023 · similarity 0.88)
- Perturbed Dynamic Time Warping: A Probabilistic Framework and Generalized Variants (ICLR 2026 · similarity 0.86)
- DTWNet: a Dynamic Time Warping Network (NeurIPS 2019 · similarity 0.85)
- TimePoint: Accelerated Time Series Alignment via Self-Supervised Keypoint and Descriptor Learning (ICML 2025 · similarity 0.84)
- Improved Learning via k-DTW: A Novel Dissimilarity Measure for Curves (ICML 2025 · similarity 0.83)
- Fortifying Time Series: DTW-Certified Robust Anomaly Detection (NeurIPS 2025 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.