When independent component analysis (ICA) is applied to color natural images, the representation it learns has spatiochromatic\r\nproperties similar to the responses of neurons in primary visual cortex. Existing models of ICA have only been applied to pixel\r\npatches. This does not take into account the space-variant nature of human vision. To address this, we use the space-variant logpolar\r\ntransformation to acquire samples fromcolor natural images, and then we apply ICA to the acquired samples.We analyze the\r\nspatiochromatic properties of the learned ICA filters. Qualitatively, the model matches the receptive field properties of neurons in\r\nprimary visual cortex, including exhibiting the same opponent-color structure and a higher density of receptive fields in the foveal\r\nregion compared to the periphery.We also adopt the ââ?¬Å?self-taught learningââ?¬Â paradigm from machine learning to assess the modelââ?¬â?¢s\r\nefficacy at active object and face classification, and the model is competitive with the best approaches in computer vision
Loading....