Research map / Video / Action recognition
DeFB: Decomposed Feature Learning for Real-Time Multi-Person Eyeblink Detection in Untrimmed In-the-Wild Videos
Jinfang Gan, Wenzheng Zeng, Yang Xiao, Xintao Zhang, Chaoyang Zheng, Ran Zhao, Ran Wang, Min Du, Zhiguo Cao
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 decomposed feature space (coarse face features plus dedicated fine-grained eye module) on top of async training (ablation) | DeFB · MPEblink | Blink-AP0.5:0.95: 12.93 -> 24.65 | Improved |
| add asynchronous training on top of decomposed features (ablation) | DeFB · MPEblink | Blink-AP0.5:0.95: 14.47 -> 24.65 | Improved |
| add both decomposed features and asynchronous training (ablation) | InstBlink-style unified model · MPEblink | Blink-AP0.5:0.95: 7.96 -> 24.65 | Improved |
| add dynamic proposal selection instead of rasterized uniform proposals (ablation) | DeFB · MPEblink | Inst-AP: 74.63 -> 76.07 | Improved |
| drop the high-resolution encoder scale for face modeling (toy experiment) (ablation) | InstBlink · MPEblink | Inst-AP: -1.34% | No clear effect |
| add separate lightweight spatio-temporal module for eyeblink (toy experiment) (ablation) | InstBlink · MPEblink | Blink-AP: +39.47% | Improved |
| decompose feature space, dropping the fine-grained scale for faces (toy experiment) (ablation) | InstBlink · MPEblink | efficiency: +14.28% | Improved |
| plug DeFB into general action detectors (ActionFormer, TriDet, DyFADet, DiGIT) | action detectors · MPEblink | Blink-AP: 1.42 -> 26.36 (Actionformer) | Improved |
| DeFB instead of InstBlink for cross-dataset transfer | unified eyeblink detectors · HUST-LEBW (trained on MPEblink) | F1: 83.45 -> 86.92 | Improved |
| DeFB instead of InstBlink on HUST-LEBW recall | unified eyeblink detectors · HUST-LEBW (trained on MPEblink) | Recall: 91.34 -> 88.98 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- DyFADet — DyFADet: Dynamic Feature Aggregation for Temporal Action Detection (ECCV 2024) · better on 1 of 1 shared result
- TriDet — TriDet: Temporal Action Detection With Relative Boundary Modeling (CVPR 2023) · better on 1 of 1 shared result
- Actionformer — ActionFormer: Localizing Moments of Actions with Transformers (ECCV 2022) · 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
- Real-Time Multi-Person Eyeblink Detection in the Wild for Untrimmed Video (CVPR 2023 · similarity 0.92)
- Detecting Attended Visual Targets in Video (CVPR 2020 · similarity 0.85)
- Gaze Target Detection by Merging Human Attention and Activity Cues (AAAI 2024 · similarity 0.84)
- Sharingan: A Transformer Architecture for Multi-Person Gaze Following (CVPR 2024 · similarity 0.84)
- Unsupervised Gaze Representation Learning from Multi-view Face Images (CVPR 2024 · similarity 0.83)
- Gaze-LLE: Gaze Target Estimation via Large-Scale Learned Encoders (CVPR 2025 · similarity 0.83)
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.
