With the growing popularity of biometrics technology in the pattern recognition field, especially identification of human has gained the attention of researchers from both academia and industry. One such type of biometric technique is Gait recognition, which is used to identify a human being based on their walking style. Generally, two types of approaches are adopted by any algorithm designed for gait recognition, namely model based and model free approaches. The key reason behind the popularity of gait recognition is that it can identify a person from a considerable distance while other biometrics has failed to do so. In this paper, the authors have conducted a survey of extant studies on gait recognition in consideration of gait recognition approaches and phases of a gait cycle. Moreover, some aspects like floor sensors, accelerometer based recognition, the influences of environmental factors, which are ignored by exiting surveys, are also covered in our survey study. The information of gait is usually obtained from different parts of silhouettes. This paper also describes different benchmark datasets for gait recognition. This study will provide firsthand knowledge to the researchers working on the gait recognition domain in any real-world field. It has been observed that work done on the gait recognition with sufficiently high accuracy is limited in comparison to research on various other biometric recognition systems and has enough potential for future research.

随着生物识别技术在模式识别领域的日益普及，特别是 人的识别已经引起了学术界和工业界研究人员的关注。一种这样的生物识别技术是步态识别，它用于根据步行方式识别人。通常，为步态识别设计的任何算法都采用两种类型的方法，即基于模型的方法和无模型的方法。步态识别之所以流行，其关键原因是它可以在相当长的距离内识别出一个人，而其他生物识别技术却无法做到这一点。在本文中，作者对步态识别的现有研究进行了调查，其中考虑了步态识别方法和步态周期的各个阶段。此外，一些方面，例如地面传感器，基于加速度计的识别，环境因素的影响（现有调查已将其忽略），我们的调查研究也涵盖了这些内容。步态信息通常是从轮廓的不同部分获得的。本文还描述了用于步态识别的不同基准数据集。这项研究将为在任何实际领域中从事步态识别领域工作的研究人员提供第一手知识。已经观察到，与对各种其他生物特征识别系统的研究相比，以足够高的准确度在步态识别上进行的工作受到限制，并且对于未来的研究具有足够的潜力。