
en
0.25
0.5
0.75
1.25
1.5
1.75
2
People Watching: Human Actions as a Cue for Single View Geometry
Published on Feb 4, 20255932 Views
We present an approach which exploits the coupling between human actions and scene geometry. We investigate the use of human pose as a cue for single-view 3D scene understanding. Our method builds up
Related categories
Presentation
People Watching: Human Actions as a Cue for Single-View Geometry00:00
Indoor Single-View 3D Geometry01:04:44
Where are the people?10:54:57
But People Are Interesting!16:20:29
People as Clutter?18:52:52
People – Cues not Clutter24:34:58
Affordances – Where can I Sit?31:05:23
Our Goal – Inverse Problem42:35:59
Our Approach (4)54:51:19
Our Approach (5)56:42:31
Our Approach (6)58:18:23
Our Approach (7)61:48:06
Detecting Human Actions (1)63:32:58
Detecting Human Actions (2)69:35:45
Additional Detectors74:11:30
DPM Detections76:28:53
Our Approach (8)79:30:56
From Poses to Functional Regions (1)81:59:26
From Poses to Functional Regions (2)92:22:49
Affordance Constraints95:43:35
Our Approach (9)97:32:50
3D Room Hypotheses100:18:54
Appearances105:30:42
Appearances Can Be Deceiving110:50:28
What If We Observe People?115:59:15
Penalties From Functional Regions (1)117:52:29
Penalties From Functional Regions (2)128:05:39
Reranking Results132:40:52
Our Approach (10)134:10:37
Estimating free space (1)135:52:33
Estimating free space (2)141:13:53
Estimating free space (3)143:22:50
Qualitative Example (1)146:30:43
Qualitative Example (2)157:01:55
Quantitative Results - Timelapses164:34:30
Single Images with People (1)171:50:23
Single Images with People (2)177:04:51
Quantitative Results – Single Image180:25:49
Our Work on Semantics193:56:57