Object recognition in a household environment performed by a service robot. The objects are recognized by shape, texture and color-based algorithms whose outputs are then combined to create an object hypothesis. This works score-based, meaning that arbitrary algorithms and recognition frameworks can be combined without changing their outputs. In contrary to similar approaches, the proposed method automatically selects which algorithms to use and how to weigh their outputs. Reference: T. Fromm, B. Staehle, W. Ertel: Robust Multi-Algorithm Object Recognition Using Machine Learning Methods. IEEE International Conference on Multisensor Fusion and Information Integration, Hamburg, Germany, 2012.
Robust Multi-Algorithm Object Recognition Using Machine Learning Methods
Experiment/Model Test
2012-01-01
26MB, 00:02:15:06
Audio-visual
Audio-visual
English
A robust multilevel segment description for multi-class object recognition
British Library Online Contents | 2015
|Automatic Selection of Object Recognition Methods using Reinforcement Learning
BASE | 2010
|Spatial Multi-object Recognition Based on Deep Learning
IEEE | 2019
|A robust machine learning structure for driving events recognition using smartphone motion sensors
Taylor & Francis Verlag | 2024
|Robust sequential view planning for object recognition using multiple cameras
British Library Online Contents | 2009
|