Research map / Detection and segmentation / Open-set and few-shot detection
Phrase Localization and Visual Relationship Detection With Comprehensive Image-Language Cues
Bryan A. Plummer, Arun Mallya, Christopher M. Cervantes, Julia Hockenmaier, Svetlana Lazebnik
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 object detectors (with box regression) (ablation) | CCA · Flickr30k Entities test | accuracy: 43.09 -> 45.29 | Improved |
| add size cue (ablation) | CCA+Det · Flickr30k Entities test | accuracy: 45.29 -> 51.45 | Improved |
| add adjective cue (ablation) | CCA+Det+Size · Flickr30k Entities test | accuracy: 51.45 -> 52.63 | Improved |
| add verb cue (ablation) | CCA+Det+Size+Adj · Flickr30k Entities test | accuracy: 52.63 -> 54.51 | Improved |
| add position cue (ablation) | CCA+Det+Size+Adj+Verbs · Flickr30k Entities test | accuracy: 54.51 -> 55.49 | Improved |
| add phrase-pair verb cues (ablation) | SPC · Flickr30k Entities test | accuracy: 55.49 -> 55.53 | No clear effect |
| add preposition pair cues (ablation) | SPC+Verbs · Flickr30k Entities test | accuracy: 55.53 -> 55.62 | No clear effect |
| add clothing and body-part pair cues (ablation) | SPC+Verbs+Preps · Flickr30k Entities test | accuracy: 55.62 -> 55.85 | No clear effect |
| add object detector cue on affected phrases (ablation) | baseline (CCA, position, size) · Flickr30k Entities, affected phrases | accuracy: 74.25 -> 75.78 | Improved |
| add adjective cue on affected phrases (ablation) | baseline (CCA, position, size) · Flickr30k Entities, affected phrases | accuracy: 57.71 -> 64.35 | Improved |
| add verb-object cue on affected phrases (ablation) | baseline (CCA, position, size) · Flickr30k Entities, affected phrases | accuracy: 40.70 -> 47.62 | Improved |
| use rank-SVM instead of direct search for cue weights (ablation) | SPC+PPC · Flickr30k Entities test | accuracy: 8% worse | Hurt |
Closest papers by meaning
- Phrase Localization Without Paired Training Examples (ICCV 2019 · similarity 0.92)
- Align2Ground: Weakly Supervised Phrase Grounding Guided by Image-Caption Alignment (ICCV 2019 · similarity 0.89)
- Revisiting Image-Language Networks for Open-Ended Phrase Detection (TPAMI 2020 · similarity 0.89)
- Visual Relationship Detection with Language Priors (ECCV 2016 · similarity 0.89)
- Flickr30k Entities: Collecting Region-to-Phrase Correspondences for Richer Image-to-Sentence Models (IJCV 2016 · similarity 0.89)
- Improved Visual Grounding through Self-Consistent Explanations (CVPR 2024 · similarity 0.87)
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.
