Learning to Segment Images Using Dynamic Feature Binding
By Michael C. Mozer, Richard S. Zemel, Marlene Behrmann, Christopher K. I. Williams. 1991
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
- J. Urgen Schmidhuber
- Yoshua Bengio
- Klaus Greff
- Sjoerd van Steenkiste
- Sergey Levine
- Trevor Darrell
- Kristen Grauman
- Ruslan Salakhutdinov
Citations in the corpus
- On the Binding Problem in Artificial Neural Networks (2020, background): “, 1995) and complex-valued RNNs that could be trained either with backpropagation (Mozer et al., 1992) or via Hebbian learning (Rao et al.”
- Toward a Biased Competition Account of Object-Based Segregation and Attention (2000, background): “Models that do not have the difficulties associated with interactive constraint satisfaction (e.g., Mozer, 1999; Mozer and Sitton, 1998; Mozer et al., 1992) may explain the mechanisms of biased competition .”
- The time dimension for scene analysis (2005, methodology): “A later study using an extended version of the backpropagation algorithm to perform a segmentation task was made by Mozer et al. [90], but again, success on only toy problems was reported.”
- Neuronal Synchrony in Complex-Valued Deep Networks (2013, methodology): “A related approach is used by Mozer et al. (1992) to model binding by synchrony in vision, performing phase-based segmentation of simple geometric contours, and by Behrmann et al. (1998) to model aspects of object-based attention.”
- Experience-Dependent Perceptual Grouping and Object-Based Attention (2002, result): “After training, MAGIC successfully segregated features of novel images into separate objects.”
- Transformation Equivariant Boltzmann Machines (2011, background): “For example in [15] the authors present a convolutional architecture where the rotational variable denotes the phase of an oscillator, relating to the theory of binding-by-synchrony.”
- Lending direction to neural networks (1995, background): “In magic (Mozer, Zemel and Behrmann, 1992; Mozer et al.,1992), directional values are used to represent alternative labels that can be assigned toimage features.”