Classier training based on synthetically generated samples
Frank Lindner, Ulrich Kreel · 2007
In most image classication systems, the amount and qual- ity of the training samples used to represent the dierent pattern classes are important factors governing the recognition performance. Hence, it is usually necessary to acquire a representative set of training samples by acquisition of data in real-world environments. Such procedures may require considerable eorts and furthermore often generate a training set which is unbalanced with respect to the number of available samples per class. In this contribution we regard classication tasks for which each real-world training sample is derived from an ideal class representa- tive which undergoes a geometric and photometric transformation. This transformation depends on system-specic inuencing quantities of the image formation process such as illumination, characteristics of the sen- sor and optical system, or camera motion. The parameters of the trans- formation model are learned from object classes for which a large number of real-world samples are available. For each individual real-world sample a set of model parameters is derived by correspondingly tting the trans- formed ideal sample to the observed sample. The obtained probability distribution of model parameters is used to generate synthetic sample sets for all regarded pattern classes. This training approach is applied to a vehicle-based vision system for trac sign recognition. Our experimen- tal evaluation on a large set of real-world test data demonstrates that the classication rates obtained for classiers trained with synthetic samples are comparable to those obtained based on real-world training data.