Phrase Detection on VRD 1.0 (test)
19.5Recall @ IoU=1.00 (All)Ours [S+A] + Language
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Ours [S+A] + LanguageSupervision=Full sup., Features=Spatial+Appearance, Language Model=From [31]2017.07 | 19.5 | — | — | 7.8 | |
| Ours [S+A]Supervision=Full sup., Features=Spatial+Appearance2017.07 | 18.1 | — | — | 8.7 | |
| Ours [S+A]Supervision=Weak sup., Features=Spatial+Appearance2017.07 | 17.4 | — | — | 7.4 | |
| Language (full)Supervision=Full sup.2017.07 | 17.3 | — | — | 5.5 | |
| Visual + LanguageSupervision=Full sup.2017.07 | 17 | — | — | 3.7 | |
| Ours [S+A] - NoisySupervision=Weak sup., Features=Spatial+Appearance, Noise=Noisy2017.07 | 16.9 | — | — | 6.7 | |
| Ours [A]Supervision=Full sup., Features=Appearance2017.07 | 16.4 | — | — | 6 | |
| Ours [S]Supervision=Full sup., Features=Spatial2017.07 | 15 | — | — | 8.7 | |
| VisualSupervision=Full sup.2017.07 | 2.6 | — | — | 1.1 | |
| Language (likelihood)Supervision=Full sup.2017.07 | 0.08 | — | — | 0.01 | |
| Visual PhrasesSupervision=Full sup.2017.07 | 0.07 | — | — | — | |
| Joint CNN2016.07 | — | 0.09 | 0.07 | — | |
| Visual Phrases2016.07 | — | 0.07 | 0.04 | — | |
| VRD - L onlyvisual_module=false, language_module=true2016.07 | — | 0.08 | 0.08 | — | |
| VRD - V + L + Kvisual_module=true, language_module=true, knowledge_kernel=true2016.07 | — | 17.03 | 16.17 | — | |
| VRD - V + L + Reg.visual_module=true, language_module=true, regularization=true2016.07 | — | 8.91 | 9.6 | — | |
| VRD - V + L onlyvisual_module=true, language_module=true2016.07 | — | 8.59 | 9.13 | — | |
| VRD - V + naive FCvisual_module=true, language_module=naive FC2016.07 | — | 6.39 | 6.65 | — | |
| VRD - V onlyvisual_module=true, language_module=false2016.07 | — | 2.61 | 2.24 | — |