Monocular estimation of 3D poses from a distance
Most 3D pose estimators only estimate egocentric coordinates where the body is centred at the origin. This is suitable for scenes with a single person but for images with interacting persons it is insufficient. We propose a monocular depth estimator for telephoto lenses to estimate 3D coordinates ce...
Elmentve itt :
Szerzők: | |
---|---|
Testületi szerző: | |
Dokumentumtípus: | Könyv része |
Megjelent: |
2018
|
Sorozat: | Conference of PhD Students in Computer Science
11 |
Kulcsszavak: | Számítástechnika |
Online Access: | http://acta.bibl.u-szeged.hu/61780 |
Tartalmi kivonat: | Most 3D pose estimators only estimate egocentric coordinates where the body is centred at the origin. This is suitable for scenes with a single person but for images with interacting persons it is insufficient. We propose a monocular depth estimator for telephoto lenses to estimate 3D coordinates centred at the camera. Our method fuses a depth map predictor and a relative 3D pose estimator by means of a 3-layer neural network. We compare the algorithm with the state-of-the-art method and show a 19% improvement. |
---|---|
Terjedelem/Fizikai jellemzők: | 114-117 |