Forensic Gait Analysis: A Tool for Justice or Oppression?
What is Forensic Gait Analysis?
Forensic gait analysis is defined as ‘the assessment and evaluation of the gait patterns and features of the person/suspect and comparing these features with the scene of crime evidence for criminal/personal identification.’ Human gait can be separated into 24 components, which make up strides, and human walking occurs in a specific pattern, including various stages, which is referred to as a ‘gait cycle.’ It mainly consists of two phases, the stance phase, and the swing phase. Analyzing it includes looking at the degree to which the height of the heel changes, and the different characteristics of the change in the distance between the feet.
The research on gait analysis first began in the early 1990s as a small laboratory group in the US started recording and analyzing the gait of a small population, and the US Department of Defense developed it into a program called HumanID.
How is Forensic Gait Analysis used in criminal courts?
Gait analysis is commonly used to identify criminals spotted on the closed-circuit television camera (CCTV) footage, and it has been used to solve a variety of criminal cases including HBT (House Break-in and Theft), robbery, sexual assaults, hit and run, shoplifting, homicides, kidnapping, and more. It is a helpful aid in the estimation of the sex of the person, estimation of body weight, as well as the type of footwear used, the use of any walking aid or support, underlying disease, disorder, a medical condition affecting the gait, etc (Badiye). Ever since gait analysis was first admitted as evidence into criminal proceedings in the UK in 2000, it has helped unravel countless criminal cases, and criminals have been convicted based on forensic gait analysis all around the world.
What are some criticisms/concerns about Forensic Gait Analysis?
Although forensic gait analysis has proven to be an efficient instrument in solving crime cases, researchers advise governmental institutions not to rely solely on its data, as its accuracy, reliability, and admissibility in the court of law are still in question.
Most essentially, gait patterns can be highly affected by various parameters, such as weather, time of the day, injuries, etc., so the investigating officer has to take several factors into account. Another limitation of gait pattern analysis is that there is not enough data to be used for gait comparison. Moreover, experts are not following any standard protocol for the analysis of gait; consequently, there are variations in the methodology used for the analysis, which creates confusion amongst the judges and jury to decide the cases based on this parameter of identification (Badiye).
Lastly, because the police have access to CCTV footage as well as the ability to instantly distinguish individuals by analyzing their gait patterns, gait analysis can work as another means of unwanted control over citizens.
How is Forensic Gait Analysis used in South Korea?
South Korea is one of the leading nations in the IT world. Korea adopted forensic gait analysis in 2014, and the first case in which gait analysis technology was taken in as sufficient evidence in a criminal case in Korea was a murder case in April 2015. At that time, a dead body was found on the banks of the Geumho River in Daegu. The murderer was spotted on the CCTV near the river, but couldn’t be identified as he covered his face with an umbrella in the blurry footage.
Experts analyzed the CCTV footage and found out that the murderer’s “walk patterns of the four videos secured by the broadcasting station all show the same characteristics. This man has a varus (so-called O-shaped) leg below his knee and has an unstable gait. He also has an unusual walk pattern with his feet pointed outward” (Baek). This result was adopted as strong evidence, and the suspect was consequently proved guilty. Ever since, gait analysis has gained a reputation as a trustworthy tool, and between 2014 and 2020, it was utilized 291 times in South Korea alone.
On January 13th, 2022, the minister of Science and Information and Communications Technology (Ministry of Science and ICT), Lim Hye-Sook, and the national police chief, Kim Chang-Ryong announced a new yearly plan. The governmental institution and the police are working together to develop AI technology to enhance public security in South Korea.
This plan includes several individual projects such as developing a chatbot for school violence, a solution to combat illegal drones, technology to enhance utilizing brain waves for allegations in criminal cases, etc. One of the projects they emphasize the most is the plan to develop forensic gait analysis further in the years to come. When the technology was first adopted, it was limited to relying heavily on professionals’ subjective experience and knowledge. Thus it lacked objectivity as the results can vary significantly depending on the professional conducting the research. In order to combat this issue, the government invested in a new software prototype, but it wasn’t long until they found frequent errors in the program interface.
As a result, in 2022, the institution is planning on including an artificial intelligence algorithm in this process. This program will have the algorithm predict the gait pattern by translating two-dimensional pictures to three-dimensional models, automatically predict the subjects’ leg joint points from the CCTV data, complete advanced analysis of changes by various gait parameters (gait speed, bag type, and weight, etc.), as well as improve the identification accuracy of the recognition model. The government plans on investing 600 million KRW (50 million US dollars) per year for two years to develop this technology.
South Korea may be in the lead in the information technology world, but its legislative actions seem unclear in comparison to other nations. Unlike the UK government, who established firm guidelines for the use of gait analysis (Code of Practice for Forensic Gait Analysis) in 2019, South Korea does not seem to have detailed regulations specific to the use of gait analysis, at least not one that is as easily accessible and thorough as the UK Code of Practice. A short phrase: “police officers and forensic agents can utilize the technology when necessary, and need to submit an analysis report afterward” is included in the National Police Agency Basic Guidelines for Scientific Investigation, but further guidelines cannot be found at the moment. Most crucially, laws regarding AI algorithms in the analysis process such as how to handle AI-produced errors, how to filter biases, and how to protect citizens’ privacy are not present. With rapidly evolving technology and various national projects being set out, South Korea needs to focus on setting the boundaries and making them clear to the public in order to establish a secure and transparent society.
So…?
Looking at South Korea’s example, it is clear that forensic gait analysis can be beneficial in helping retroactively identify criminals, but adopting this technology combined with AI would require articulate restrictions and laws. Gait analysis is not as well-known as other AI-based technology such as facial recognition technology. Facial recognition technology raised many concerns and scrutiny all over the world regarding privacy and freedom of speech, and there are laws and moratoriums being made in various nations to prevent social injustice reproduced by AI. Yet, the use of gait analysis is not something that has been debated frequently but is definitely something that could initiate similar issues.
The fundamental questions that can be raised at this point are — Will AI really solve the pre-existing concerns about the reliability and safety of the technology or would it only be regenerating biases? Do we have sufficient guidelines/laws?
All of these questions really come down to the final discussion: Are we ready?
Sources
Badiye, A. (2021, November 14). Forensic gait analysis. StatPearls [Internet]. Retrieved March 6, 2022, from https://www.ncbi.nlm.nih.gov/books/NBK557684/
Iwama, H. (2013). Gait Verification System for Criminal Invetigation. Retrieved March 6, 2022, from http://www.am.sanken.osaka-u.ac.jp/~makihara/pdf/cva_2013Oct_iwama.pdf.
Seckiner, D. (2019). Forensic gait analysis — Morphometric assessment from surveillance footage. ScienceDirect. Retrieved March 6, 2022, from https://www.google.com/url?sa=i&url=https%3A%2F%2Fwww.sciencedirect.com%2Fscience%2Farticle%2Fpii%2FS0379073819300167&psig=AOvVaw1MhLyU7Z7FI8Fnqy-YzYZv&ust=1646662733077000&source=images&cd=vfe&ved=0CAsQjRxqFwoTCPDSlO3WsfYCFQAAAAAdAAAAABAD.
The Royal Society. (2017, November). Royal Society Forensic Gait Analysis Primer for Courts. Retrieved March 6, 2022, from https://royalsociety.org/~/media/about-us/programmes/science-and-law/royal-society-forensic-gait-analysis-primer-for-courts.pdf
과학치안 공공연구성과 실용화 촉진 — 사업공고. Msit.go.kr. (n.d.). Retrieved March 6, 2022, from https://www.msit.go.kr/bbs/view.do?sCode=user&mId=129&mPid=128&bbsSeqNo=100&nttSeqNo=3177498
Baek, J. (2017). [과학] 용의자 식별의 진일보, 보행인식. 대학원신문. Retrieved March 6, 2022, from http://gspress.cauon.net/news/articleView.html?idxno=21451
