Abstract: Primates are very good at recognizing objects and at the same time assessing their viewing angle and relative position. In such tasks they outperform existing computer vision systems. We present a model for the unsupervised learning of object identities and codes for position and viewing angle (including depth rotation). The model is based on a hierarchy of Slow Feature Analysis (SFA) modules, which were shown to be a good model for complex cells in the early visual system. Our model is trained and tested with visual input data generated from virtual 3D-objects. To show that the model extracts useful information we use linear regressions (for position and viewing angle) or simple classifiers (for object identity). This supervised step extracts the feature codes with high invariance under the transformations.