Log-Linear Mixtures for Object Recognition
Tobias Weyand, Thomas Deselaers and Hermann Ney
In: BMVC 2009, 7-10 Sep 2009, London, UK.
We present log-linear mixture models as a fully discriminative approach to object category recognition which can, analogously to kernelised models, represent non-linear decision boundaries. It is shown that this model is the discriminative counterpart to Gaussian mixtures and that either one can be transformed into the respective other. However, the proposed model is easier to extend toward fusing multiple cues and numerically more stable to train and to evaluate. We experimentally evaluate our model on the PASCAL VOC 2006 data and the results compare favourably well to the state-of-the-art despite the model consisting of an order of magnitude fewer parameters, which suggests excellent generalisation capabilities.