Biological Cybernetics - Call for papers: Biological Cybernetics Special Issue: What can Computer Vision learn from Visual Neuroscience?
Computer vision still struggles to solve many problems as quickly and accurately as our brain does, despite its tremendous progress using deep learning over the last decade. Current challenges include geometry understanding, dynamic scene understanding, few shot learning of novel objects, anomaly and counterfeit detection, stable reproduction of imaginary visuals, and more. The brain achieves all these on a low computational cost and power budget (i.e., tens of watts), whereas the state-of-the-art computer vision models require multiple graphics processors consuming kilowatts of power. Hence, new inspirations from the brain could make computer vision more efficient, robust, and capable of continuous learning and adaptation. Although visual neuroscience literature provides knowledge about neuronal selectivities and early processing of visual inputs, computational modelling of neuronal and network behaviours with some degree of biological constraints and characteristics is an important tool to address many of the challenges facing computer vision. With new neuroscientific insights about deep cortical vision processing using the latest stimulation and recording technologies, it is important that we strive to understand the underlying computations and learning mechanisms, so that artificial vision systems with similar efficiency can be developed. This special issue invites original research and review articles related to topics in biological vision that can potentially benefit computer vision systems. The following is a non-exhaustive list of topics in biological vision that can also benefit computer vision systems.
● Active vision’s role in visual search, scene understanding, social interactions, etc.
● Learning in the visual system. Learning in biology is continual, few-shot, and adversarially robust.
● The roles of recurrent and top-down connections in the visual cortex.
● Spike based spatiotemporal processing and its implications for neuromorphic vision.
● Motion perception in dynamic environments.
● Neural coding schemes in the visual system (e.g., sparse coding, predictive coding, and temporal coding.)
● The roles of attention mechanisms in biological vision.
How to submit
Please submit your manuscript using the journal online submission system (this opens in a new tab) following the Biological Cybernetics submission guidelines (this opens in a new tab).
The special issue will open for submissions from 1st June with a closing deadline of 30 September 2022.
Guest Editors
Kexin Chen - Department of Cognitive Sciences, University of California, Irvine
Hirak J. Kashyap - Department of Computer Science, University of California, Irvine
Jeffrey L. Krichmar - Department of Cognitive Sciences, Department of Computer Science, University of California, Irvine
Xiumin Li - College of Automation, Chongqing University, Chongqing, China