Research map / Video / Action recognition
Face-Guided Sentiment Boundary Enhancement for Weakly-Supervised Temporal Sentiment Localization
Cailing Han, Zhangbin Li, Jinxing Zhou, Wei Qian, Jingjing Hu, Yanghao Zhou, Zhangling Duan, Dan Guo
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 |
|---|---|---|---|
| add facial features to audio-visual input (ablation) | FSENet · TSL300 | average mAP: 19.41 -> 21.45 | Improved |
| add GSP branch (ablation) | FSENet · TSL300 | average mAP: 20.22 -> 21.45 | Improved |
| add facial guidance to both FCI stages (ablation) | FSENet · TSL300 | average mAP: 18.42 -> 21.45 | Improved |
| add facial guidance only in FCI second stage (ablation) | FSENet · TSL300 | average mAP: 18.42 -> 18.78 | No clear effect |
| add boundary-aware step pseudo-label to threshold pseudo-label (BSPG) (ablation) | FSENet · TSL300 | average mAP: 20.29 -> 21.45 | Improved |
| add confidence-threshold pseudo-labels over raw predictions (ablation) | FSENet · TSL300 | average mAP: 16.92 -> 20.29 | Improved |
| add frame loss L_frame (ablation) | FSENet · TSL300 | average mAP: 20.02 -> 21.45 | Improved |
| add global frame loss (ablation) | FSENet · TSL300 | average mAP: 20.65 -> 21.45 | Improved |
| add semantic contrast loss L_sc (ablation) | FSENet · TSL300 | average mAP: 20.94 -> 21.45 | Improved |
| use cosine similarity instead of dot product in PSSC (ablation) | FSENet · TSL300 | average mAP: 21.00 -> 21.45 | Improved |
| use cosine similarity instead of L1 in PSSC (ablation) | FSENet · TSL300 | average mAP: 20.05 -> 21.45 | Improved |
| add BSPG pseudo-labels to LLaMA-2-7B LoRA (ablation) | LLaMA-2-7B · TSL300 | average mAP: 9.97 -> 11.75 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DDG-Net — DDG-Net: Discriminability-Driven Graph Network for Weakly-supervised Temporal Action Localization (ICCV 2023) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- HR-Pro: Point-Supervised Temporal Action Localization via Hierarchical Reliability Propagation (AAAI 2024 · similarity 0.83)
- Realigning Confidence with Temporal Saliency Information for Point-Level Weakly-Supervised Temporal Action Localization (CVPR 2024 · similarity 0.83)
- Seek Common Ground While Reserving Differences: Semi-Supervised Image-Text Sentiment Recognition (CVPR 2025 · similarity 0.83)
- Boundary-Aware Temporal Dynamic Pseudo-Supervision Pairs Generation for Zero-Shot Natural Language Video Localization (AAAI 2025 · similarity 0.83)
- MART: Masked Affective RepresenTation Learning via Masked Temporal Distribution Distillation (CVPR 2024 · similarity 0.82)
- Rethinking Pseudo-Label Guided Learning for Weakly Supervised Temporal Action Localization from the Perspective of Noise Correction (AAAI 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.
