Pattern Recognition is the official journal of the Pattern Recognition Society. The Society was formed to fill a need for information exchange among research workers in the pattern recognition field. Up to now, we ''pattern-recognitionophiles'' have been tagging along in computer science, information theory, optical processing techniques, and other miscellaneous fields. Because this work in pattern recognition presently appears in widely spread articles and as isolated lectures in conferences in many diverse areas, the purpose of the journal Pattern Recognition is to give all of us an opportunity to get together in one place to publish our work. The journal will thereby expedite communication among research scientists interested in pattern recognition.

We consider pattern recognition in the broad sense, and we assume that the journal will be read by people with a common interest in pattern recognition but from many diverse backgrounds. These include biometrics, target recognition, biological taxonomy, meteorology, space science, classification methods, character recognition, image processing, industrial applications, neural computing, and many others.

The publication policy is to publish (1) new original articles that have been appropriately reviewed by competent scientific people, (2) reviews of developments in the field, and (3) pedagogical papers covering specific areas of interest in pattern recognition. Various special issues will be organized from time to time on current topics of interest to Pattern Recognition.

Special Issue on Practical Deep Learning in the Wild
Submission Date: 2022-11-01

Deep learning has achieved significant success for pattern recognition in multiple fields, including computer vision, natural language processing, and acoustics. Equipped with the powerful deep neural networks, an increasing number of applications such as autonomous driving arise. However, there still exist several challenges that hinder the broader practical use of deep learning models in the real-world scenario. First, many scenarios have strict requirements for the inference latency of neural networks, especially on the edge and mobile devices with limited computation resources (e.g., camera or mobile phones). Second, in some rare but important events such as flame detection, it is hard for a deep model to learn and generalize well due to the data sparsity. However, it is infeasible to obtain enough training data, which indicates the significance of approaches to improving the data efficiency. Third, even if we produce an efficient neural network with limited data, its robustness is still being challenged since the inputs are often perturbed with various noises in the wild. The unreliability and vulnerability brought by the noises will pose a safety risk to the actual deployment. Taking all these aspects into consideration, there is still a long way to go for practical deep learning in the wild.

Therefore, the practical deep learning in the wild should take the following aspects into consideration including computation efficiency, data efficiency and robustness on various noises. In this special issue, we aim to bring researchers from these areas together to address the open and tough challenges and further build practical deep learning applications and systems in the wild.

Topics:

    Techniques to improve computation efficiency, including but not limited to neural network compression, efficient neural network architecture design, high performance hardware design and software implementation optimization
    Algorithms and theories for data efficiency that can learn and generalize well with few data such as zero-shot/few-shot learning, continual learning, learning with human in the loop and so on.
    Techniques, systems, or frameworks to build trustworthy and robust models towards diverse noises in the wild (e.g., natural noises, adversarial noises, system noises, etc.)
    Benchmark for practical deep learning system and algorithms
    Theoretical or empirical studies on the relationship among the three factors of practical deep learning: computation efficiency, data efficiency and robustness.

Time schedule:

    Call for papers: August 1, 2022
    Paper submission deadline: November 1, 2022
    First review: January 1, 2023
    Revisions due: February 1, 2023
    Final decisions: March 1, 2023
    Tentative publication date: May 1, 2023

Guest Editors:

    Xianglong Liu, Beihang University. xlliu@buaa.edu.cn
    Wanli Ouyang, the University of Sydney. wanli.ouyang@sydney.edu.au
    Nicu Sebe, University of Trento. sebe@disi.unitn.it
    Dawn Song, UC Berkeley. dawnsong@gmail.com
    Aishan Liu, Beihang University. liuaishan@buaa.edu.cn
    Ruihao Gong, SenseTime. gongruihao@sensetime.com