About this Research Topic
Visual data is one of the most important information sources for human, from which we are able to obtain well recognition of the world. Therefore, computer vision is one of the most important research areas in artificial intelligence, which aims to enable the intelligent system to understand and analysis the visual information as human. Thanks to the recent progress of deep learning, the performances of current computer vision techniques are boosted, even outperforming human in certain tasks. However, most of the existing deep learning techniques for computer vision achieve promising performance only on carefully collected datasets, which limits their implementation on real-world environments that with large amounts of variants and noises.
The robust solutions for computer vision are urgently demanded for the implementations on unstable scenarios, such as those with large variations, serious noises or extreme cases. For example, few/zero-shot learning technique strive to learn to predict from limited, even not any, available samples, which increases the generalizability of the intelligent system, thus able to preserve its performance under unstable circumstances; noisy data learning techniques designed for learning to understand and analysis the data with large amount of noises, which increases the system robustness. Other techniques, such as explainable learning, out-of-distribution data understanding and analysis, domain adaptation and adversarial learning, are also essential for the development of robust solutions.
The goal of this Research Topic is to solicit novel, high-impact, high-quality and original papers that aims to propose robust deep learning solutions for computer vision tasks. We are interested in submissions including but not limited to the following topics:
• Theories, Models and Datasets for Robust Computer Vision Techniques;
• Noisy Data Learning in Computer Vision;
• Explainable Learning in Computer Vision;
• Few/Zero Shot Learning in Computer Vision;
• Out-of-distribution Data Understanding and Analysis in Computer Vision;
• Domain Adaptation in Computer Vision;
• Transfer Learning in Computer Vision;
• Adversarial Learning in Computer Vision；
• Data Augmentation in Computer Vision.
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.