People oftenfeel that they can identify a familiar person from afar simply by recognizingthe way the person walks. This common experience, combined with recent interestbiometrics, has lead to the development of gait recognition as a from of biometricidenti?cation. Gait (walk) pattern has several attractive properties as a softbiometric trait. From a surveillance perspective, gait pattern biometrics is appealing in that it can be performedat adistance without requiringbody-invasive equipment or subject cooperation. Ref 1 Many research groupsinvestigate the discrimination power of gait pattern and develop models thatare applied to the automatic recognition of walking people from MoCap data.
A number of MoCap-based gait recognition methodshave been introduced in the past few years and newones continue to emerge. In order to move forward withthis competitive research,it is necessary to comparetheir innovative approaches with thestate-of-the-art and evaluatethem against established evaluation metrics on a benchmark database. As a biometric, gait has several attractive properties. Acquisition ofimages portraying an individual’s gait can be done easily in public areas, withsimple instrumentation, and does not require the cooperation or even awarenessof the individual under observation. In fact, it seems that it is thepossibility that a subject may not be aware of the surveillance andidenti?cation that raises public concerns about gait biometrics 2. There arealso several confounding properties of gait as a biometric.
Unlike ?ngerprints, we do not know the extent to which an individual’s gait is unique.Furthermore, there are several factors, other than the individual, that causevariations in gait, including footwear, terrain, fatigue, and injury. Thispaper gives an overview of the factors that affect both human and machinerecognition of gaits, data used in gait and motion analysis, evaluationmethods, existing gait and quasi gait recognition systems, and uses of gaitanalysis beyond biometric identification.