7th International Conference on Learning Representations (, 2019). ImageNet-trained CNNs are biased towards texture increasing shape bias improves accuracy and robustness. Neocognitron: a self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position.
Backpropagation applied to handwritten zip code recognition. Optimal sensorimotor integration in recurrent cortical networks: a neural implementation of Kalman filters. Bayesian Filtering and Smoothing (Cambridge Univ. Object persistence in philosophy and psychology. On the binding problem in artificial neural networks. A perspective on objects and systematic generalization in model-based RL. Whatever next? Predictive brains, situated agents, and the future of cognitive science. Indirect Perception (The MIT Press, 1997).Ĭlark, A. Les Complements Amodaux des Structures Perceptives (Institut de psychologie de l’Université de Louvain, 1964). Natural image statistics and neural representation. Experiments in the visual perception of texture. In Human Vision and Electronic Imaging VI (eds. On seeing stuff: the perception of materials by humans and machines. The Construction of Reality in the Child (Basic Books, 1954).Īdelson, E. Vision as Bayesian inference: analysis by synthesis? Trends Cogn. Handbuch der Physiologischen Optik (Voss, 1867). The cognitive literature provides a starting point for the development of new experimental tasks that reveal mechanisms of human object perception and serve as benchmarks driving the development of deep neural network models that will put the object into object recognition. Here, we review related work in both fields and examine how these fields can help each other. By contrast, deep neural network models of visual object recognition remain largely tethered to sensory input, despite achieving human-level performance at labelling objects.
Human behavioural studies have documented how object representations emerge through grouping, amodal completion, proto-objects and object files. Object representations emancipate perception from the sensory input, enabling us to keep in mind that which is out of sight and to use perceptual content as a basis for action and symbolic cognition. Human visual perception carves a scene at its physical joints, decomposing the world into objects, which are selectively attended, tracked and predicted as we engage our surroundings.