Research map / Detection and segmentation / Weakly supervised detection and segmentation
Weakly-Supervised Temporal Localization via Occurrence Count Learning
Julien Schroeter, Kirill Sidorov, David Marshall
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| Train with occurrence counts instead of localization labels | LoCo drum transcription · D-DTD Eval Random | F1: 97.5 -> 94.5 | Hurt |
| Train with occurrence counts instead of localization labels | LoCo drum transcription · D-DTD Eval Subset | F1: 92.8 -> 87.1 | Hurt |
| Train with occurrence counts instead of localization labels | LoCo drum transcription · D-DTP Eval Random | F1: 90.0 -> 88.9 | Hurt |
| Train with occurrence counts instead of localization labels | LoCo drum transcription · D-DTP Eval Subset | F1: 79.4 -> 78.1 | Hurt |
| Train using occurrence counts instead of onset labels | LoCo piano onset detection · MAPS | F1: 82.29 -> 71.99 | Hurt |
| Train with counts for weakly supervised digit localization | LoCo digit detection · MNIST synthetic digit images | single digit recognition accuracy: 98.51% -> 99.12% | Improved |
| Tighten the event matching tolerance | LoCo drum transcription · D-DTD Eval Random, hi-hats | F1: 97.1% -> 96.3% | Hurt |
Closest papers by meaning
- Learning Precise Temporal Point Event Detection with Misaligned Labels (AAAI 2021 · similarity 0.91)
- Temporal Action Localization in the Deep Learning Era: A Survey (TPAMI 2023 · similarity 0.88)
- Weakly-Supervised Action Localization With Background Modeling (ICCV 2019 · similarity 0.86)
- Adversarial Background-Aware Loss for Weakly-supervised Temporal Activity Localization (ECCV 2020 · similarity 0.86)
- Temporal Action Localization by Structured Maximal Sums (CVPR 2017 · similarity 0.86)
- Decompose the Sounds and Pixels, Recompose the Events (AAAI 2022 · similarity 0.85)
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.
