Code/Software
Network simulator for spiking neurons
The Event-Related Neural Simulation Tool (ERNST) is a simulator of networks of spiking neurons that makes use of highly efficient event-related numerical techniques. Rather than advancing the state of the network by one time-step (of fixed or variable length), for every neuron, the simulator directly ’jumps’ to the next spike in one step. This is particularly useful for large networks with sparse activity and connectivity. The simulator is written in JAVA and does not require compilation. It is freely available at https://sourceforge.net/p/ernst/wiki/Home. For more details see: S. Mihalas and E. Niebur. A Generalized Linear Integrate-And-Fire Neural Model Produces Diverse Spiking Behavior. Neural Computation, 21(3):704–18, 2009. Stefan Mihalas, Yi Dong, Rüdiger von der Heydt, and Ernst Niebur. Event-Related Simulation of Neural Processing in Complex Visual Scenes. In 45th Annual Conference on Information Sciences and Systems IEEE-CISS 2011.
Proto-object based saliency map model
Feature-based attentional selection based on the saliency map model (see E. Niebur and C. Koch. Control of Selective Visual Attention: Modeling the “Where” Pathway. In D. S Touretzky, M. C. Mozer, and M. E. Hasselmo, editors, Advances in Neural Information Processing Systems, volume 8, pages 802–808. MIT Press, Cambridge, MA, 1996 and L. Itti, C. Koch, and E. Niebur. A model of saliency-based fast visual attention for rapid scene analysis. IEEE Transactions on Pattern Analysis and Machine Intelligence, 20(11):1254–1259,1998) is surprisingly successful in predicting human behavior. However, humans can also use more powerful representations in which they organize visual scenes in terms of objects. A computational model that makes of these mechanisms is described by A. F. Russell, S Mihalas, R. von der Heydt, E. Niebur, and R. Etienne- Cummings. A model of proto-object-based saliency. Vision Research, 94:1–15, 2014. The code for this model is available at https://github.com/csmslab/russell-proto-object. A related model that also includes a motion channel is described by Jamal Molin, Ralph Etienne-Cummings, and Ernst Niebur. How is Motion Integrated into a Proto-Object Based Visual Saliency Model? In 49th Annual Conference on Information Sciences and Systems IEEE-CISS, 2015. The code is available at https://github.com/csmslab/dynamic-proto-object-saliency While many neurons in the primary visual cortex are sensitive to small patches of oriented features, higher cortical areas integrate them into contours that play an important role in defining objects. A model describing how these objects are segregated from each other and the background is described in Brian Hu and Ernst Niebur. A recurrent neural model for proto-object-based contour integration and figure-ground segregation. Journal of computational neuroscience, 43(3):227–242, 2017.