Object Recognition with Multiple Feature Types

A system for person-independent hand posture recognition against complex backgrounds

“…The introduction of additional feature types gave significant improvements over just using Gabor jets. In another experiment [18], the advantages of using multiple feature types were even more pronounced: For the analysis of cluttered scenes of occluding objects against complex backgrounds, the recognition rate climbed from 10.0 percent of correctly analyzed scenes to 62.9 percent due to the use of compound jets instead of only Gabor jets. There, however, the success was more dependent on the precise weighting between the feature types.…”

“…While earlier versions of EGM have only worked with a single shape or texture feature type such as Gabor or Mallat filters, we have recently extended it for handling multiple feature types [18]. There are several ways of extending EGM for multiple feature types.…”

Semantic Modelling of Space

“…Various cue integration methods have been proposed in the robotics and machine learning community [71,84,10,85,86,87]. These approaches can be described according to various criteria.…”

Cue integration through discriminative accumulation

“…This implies longer learning and recognition times, greater memory requirements and possibly curse of dimensionality effects. Another strategy is to use integration schemes [18,26,30]. Here, the pattern recognition literature offers a vast choice, but one of the most popular methods in object recognition is the voting scheme [15,6,12].…”