Acquiring Visual Classifiers from Human Imagination

Figure 1: Although all image patches in (a) are just noise, when we show thousands of them to online workers and ask them to find ones that look like cars, a car emerges in the average, shown in (c). This noise-driven method is a well known approach in human psychophysics that extracts the decision function that the human visual system uses for recognition. We are adopting this method to build object recognition systems that use classifiers printed from the human mind.