Research map / Detection and segmentation / CNN features
End-To-End Face Detection and Cast Grouping in Movies Using Erdos-Renyi Clustering
SouYoung Jin, Hang Su, Chris Stauffer, Erik Learned-Miller
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 |
|---|---|---|---|
| use rank-1 counts instead of L2 distance for link-based clustering | link-based clustering · Hannah | F0.5: .2581 -> .6813 | Improved |
| use rank-1 counts instead of L2 distance for link-based clustering | link-based clustering · Buffy s05 (avg of 6 episodes) | F0.5: .3931 -> .6299 | Improved |
| use rank-1 counts instead of L2 distance for link-based clustering | link-based clustering · BBT s01 (avg of 6 episodes) | F0.5: .7365 -> .7828 | Improved |
| use rank-1 counts instead of L2 distance for link-based clustering | link-based clustering · LFW | pairwise F0.5: .8498 -> .8943 | Improved |
| use rank-1 counts instead of L2 in verification at FPR 1E-9 | verification · LFW all-pairs verification | recall: 0.0068 -> 0.0252 | Improved |
| use rank-1 counts instead of L2 in verification at FPR 1E-6 | verification · LFW all-pairs verification | recall: 0.1279 -> 0.1872 | Improved |
| use rank-1 counts instead of rank-order distance at FPR 1E-3 | verification · LFW all-pairs verification | recall: 0.8719 -> 0.8222 | Hurt |
| use rank-1 counts instead of template adaptation | link-based clustering · LFW | pairwise F0.5: .3735 -> .8943 | Improved |
| use automatically estimated threshold instead of oracle threshold for rank-1 counts | Rank-1 Count · Hannah | F0.5: .6813 -> .6436 | Hurt |
| use automatically estimated threshold instead of oracle threshold for rank-1 counts | Rank-1 Count · Buffy s05 (avg of 6 episodes) | F0.5: .6299 -> .5661 | Hurt |
| use automatically estimated threshold instead of oracle threshold for rank-1 counts | Rank-1 Count · LFW | pairwise F0.5: .8943 -> .8532 | Hurt |
| use automatically estimated threshold instead of oracle threshold for rank-1 counts | Rank-1 Count · BBT s01 (avg of 6 episodes) | F0.5: .7828 -> .7728 | No clear effect |
Closest papers by meaning
- VideoClusterNet: Self-Supervised and Adaptive Face Clustering for Videos (ECCV 2024 · similarity 0.91)
- Clustering Millions of Faces by Identity (TPAMI 2017 · similarity 0.91)
- Video Face Clustering With Unknown Number of Clusters (ICCV 2019 · similarity 0.89)
- Linkage Based Face Clustering via Graph Convolution Network (CVPR 2019 · similarity 0.88)
- Structure-Aware Face Clustering on a Large-Scale Graph With 107 Nodes (CVPR 2021 · similarity 0.88)
- PairDETR : Joint Detection and Association of Human Bodies and Faces (CVPR 2024 · 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.
