In July 2000, forensic history was made when a jewellery thief was prosecuted with the help of forensic evidence. But this was not through DNA analysis or fingerprint comparison, as one might expect, but by studying the perpetrator’s gait. That is, how the individual moved. When podiatrist Haydn Kelly was called upon to offer his professional opinion in the case of a suspected jewellery thief, the expert was able to identify the suspect from video footage through nothing but the study of the individual’s walking mechanics, ultimately declaring that he believed the perpetrator captured in the video was the same as the person under suspicion.
Forensic gait analysis refers to the identification of an individual based on the characteristic features of the way in which they move. This sub-section of forensic podiatry (the study of evidence relating to the human foot) is based on the theory that people have unique movement characteristics as they move their body and limbs.
These characteristics may be subtle, for instance the angle of the feet throughout movement, or they may be fairly distinguishing, such as the condition of genu varum (bow-leggedness), which led to burglar John Rigg’s identification and subsequent conviction in 2008.
An individual’s gait can be affected by a range of factors, including sex, weight, height, age, but also taking into account external factors such as terrain, clothing and if they are carrying anything. Gait is a behavioural biometric and thus is subject to change based on a person’s emotional state or whether they are, for instance, in a rush to get somewhere quickly or under the influence of drugs or alcohol. It is worth nothing that the term ‘gait’ is not synonymous with walking, but refers to locomotion in general, whether that be walking, running, crawling, and so on. Furthermore, a person’s gait does not only relate to the movement of their feet and legs, as one might expect, but can also encompass their upper body movements, namely their arms and torso.
The repetitive gait cycle is said to be divided into two distinct phases; the stance phase, during which the lower limbs are providing support for the body by being in contact with the ground, and the swing phase, at which point the foot is no longer in contact with the ground. It is the spatial and temporal measurements throughout these phases that can be of great use to the gait analysis expert.
The Forensic Analysis of Gait
With the proliferation of closed circuit television (CCTV) cameras worldwide, it is becoming increasingly commonplace that a perpetrator is caught on camera in the act of committing of a crime. However this footage may not always be beneficial in identifying the subject based on their face. Perhaps the perpetrator has taken steps to conceal his or her identity or, more likely, the footage is simply not good enough for identification. A vast array of factors can affect the usefulness of CCTV footage, including camera quality, distance, lighting, environmental conditions and the angle of the recording. Thus all investigators may be left with is the blurry figure of their perpetrator. Not especially useful for identification purposes one might think.
So how can an expert use the study of a person’s gait to identify an individual? Well, it’s a little more complex than watching someone walk and deciding they look the same as a perpetrator captured on CCTV.
Identification in gait analysis is a comparative process – the footage of a perpetrator is compared with footage of a suspect walking and the similarities and differences are assessed. This analysis may be qualitative, involving the visual evaluation of movement by an expert, or quantitative, a far more complex process involving the collection, interpretation and comparison of numerical data relating to the movements.
The analysis of video evidence of a suspect typically involves the painstaking inspection of each frame of the video, identifying the slightest characteristic movement of the legs, arms, shoulders and head. A vast array of parameters will be studied throughout the analysis, including the length of steps and stride, speed and rhythm of walking, and finer points such as the angles of feet and hips as the individual moves. The analysis will additionally extend to any possible affecting factors, such as whether the perpetrator was carrying something substantial in the footage, such as a hefty bag of stolen goods, and how this might alter his or her gait. Any affecting factors such as these must be taken into account when recording footage of a suspect for comparison purposes. Footage of a person walking in light clothing and empty-handed is hardly a suitable comparison for a video of the perpetrator in a bulky coat and carrying a heavy rucksack.
Although gait analysis is a relatively new approach in forensic science, the study of a person’s unique patterns of movement is far from novel. Gait analysis has been utilised in a medical setting for decades, focusing on the study of abnormalities in a person’s gait for the purpose of medical research and treatment. Similarly, gait has been studied in sports to aid athletes in perfecting their running technique, for instance. This history has allowed for the development of well-established procedures in studying and comparing the movement of individuals.
Unsurprisingly, forensic gait analysis is not without its caveats, and experts in the field hasten to concede that it is not an indisputable field of study. Whereas DNA analysis and fingerprint comparison can, to an extent, provide clear evidence as to a connection, gait analysis is best utilised as a guideline as to the possibility of an identification being made. Although gait analysis is well-established in a clinical setting, the evidence available in a forensic setting is typically of relatively poor quality in comparison to that available in medical studies, thus footage may not actually be suitable for gait analysis in the first place.
The analysis of gait conducted by a human expert can obviously introduce the potential for human error and subjectivity, even calling the conclusions reached by experts into question, especially considering the lack of quality control in this field of study. A 2013 study concluded that people who were experienced in visual gait analysis were only able to consistently identify individuals based on their gait 71% of the time (Birch et al, 2013), indicating the scope for error.
Despite the obvious downfalls of gait analysis conducted by humans, the possibilities of gait as a form of biometric identification has sparked a great deal of research into the development of automatic gait recognition technology. Computer-based systems have already been developed that examine video frames and separate the silhouette of the individual in question from the background, then recording that individual’s movement. This, and similar systems, in theory allows for the detection and tracking of humans by using computer systems to store an individual’s ‘gait signature’ and later identify that person when they are next caught on camera. The development and use of automated gait recognition systems would reduce the scope for human error, greatly increase the speed of analysis, and allow for real-time identification based on gait. If sufficiently developed, such technology could even form the basis of new security systems, only allowing access to the unique gait of particular individuals, much like retinal or fingerprint scanners. Though gait analysis technology may still be in its infancy, the scope of potential applications is great.
BBC News. How can you identify a criminal by the way they walk? [online] Available: http://news.bbc.co.uk/1/hi/magazine/7348164.stm
Birch, I et al. Terminology and forensic gait analysis. Sci Justice. 2015(55), pp. 279-284.
Birch, I et al. The development of a tool for assessing the quality of closed circuit camera footage for use in forensic gait analysis. J For Legal Med. 2013(20), pp. 915-917.
Birch, I et al. The identification of individuals by observational gait analysis using closed circuit television footage. Sci Justice. 2013(53), pp. 339-342.
DiMaggio, J. A. Vernon, W. 2011. Forensic Podiatry: Principles and Methods. New York: Springer Science & Business Media.
Forensic Magazine. Considerations of Gait at Crime Scenes. [online] Available: http://www.forensicmag.com/articles/2014/08/considerations-gait-crime-scenes
Jasuja, O. P. Manjula. Estimation of stature from footstep length. Forensic Sci Int. 1993(61), pp. 1-5.
New Scientist. Cameras Know You by Your Walk. [online] Available: https://www.newscientist.com/article/mg21528835-600-cameras-know-you-by-your-walk
Stevenage, S. V et al. Visual analysis of gait as a cue to identity. Appl Cognitive Psych. 1999(13), pp. 513-526.