About this Research Topic
The brain-inspired visual computing refers to seeking inspiration from various aspects of the neural structure, cognitive mechanisms, behavioral characteristics, and other dimensions of the biological brain. It incorporates advanced findings from neuroscience, cognitive science, and psychology to propose new visual computing models and methods. These aim to overcome current limitations in models and methods, ultimately enhancing the performance of visual computing in terms of accuracy, robustness, adaptability, generalization, and interpretability. The fusion of neuroscience and visual computing can be approached from two directions: one involves modeling visual computing based on insights from brain science mechanisms, while the other entails analyzing and drawing inspiration from existing neural network models that serve visual computing.
The problem that this Research Topic seeks to address is the gap between conventional computer vision systems and human visual perception. Recent advances in neuroscience, cognitive psychology, and machine learning have shed light on the intricate mechanisms governing human vision. Deep neural networks, convolutional neural networks, and recurrent architectures have been inspired by the hierarchical organization of the visual cortex. Transfer learning, attention mechanisms, and unsupervised learning techniques have also shown promise in addressing computer vision challenges. However, there remains a need to integrate these advances cohesively to create a comprehensive brain-inspired framework for computer vision.
To this end, this Research Topic will encompass a wide range of themes related to brain-inspired computer vision, including but not limited to the following:
- Hierarchical Visual Processing: Explore the hierarchical organization of the human visual system and design neural network architectures that mimic these structures.
- Attention Mechanisms: Investigate attention models and their application in visual attention, saliency detection, and object recognition tasks.
- Transfer Learning: Examine transfer learning techniques that allow the adaptation of pre-trained models to new visual tasks, leveraging the knowledge acquired from previously learned data.
- Unsupervised and Self-supervised Learning: Explore unsupervised learning paradigms that can reduce the dependency on large labeled datasets, and self-supervised techniques that encourage models to learn from their own representations.
- Cognitive Neuroscience Insights: Integrate findings from cognitive neuroscience to inform the design of more biologically plausible computer vision systems.
- Ethical and Philosophical Considerations: Discuss the ethical implications and philosophical aspects of creating brain-inspired computer vision systems, including privacy, bias, and the nature of perception.
We invite contributions in the form of Original Research articles, Review papers, Methods papers, and Opinion articles that push the boundaries of brain-inspired computer vision. Researchers are encouraged to explore novel ideas and interdisciplinary approaches that bridge the gap between artificial and biological vision, ultimately paving the way for a new era in computer vision research.
Keywords: Cognitive Sciences; Neural Computing; Brain and Mind Inspired Intelligence; Computer Vision; Neuroscience
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.