Gait recognition system: deep dive into this future tech
A person's gait is as unique as their voice timbre. Leveraging this knowledge, gait recognition technology has been developed using Machine Learning (ML) algorithms. ML-based systems can identify a person from an image even if their face is out of view, turned away from the camera, or concealed behind a mask.
The system analyzes the silhouette, height, speed, and walking characteristics to identify individuals from a database. This technique is more convenient than retinal scanners or fingerprints in public places as it is unobtrusive. Moreover, gait recognition is unlikely to be fooled — every person's gait has no duplicates.
Table of Contents
- What is gait recognition
- Gait recognition algorithm
- How does gait recognition work
- Capturing gait data
- Silhouette segmentation
- Contour detection
- Feature extraction and classification
- What do gait recognition systems analyze
- How gait recognition is used
- How accurate are gait recognition systems?
- Advantages of gait recognition
- Disadvantages of gait recognition
- Summary
- FAQ
What is gait recognition
Gait is an important indicator that is used in behavioral biometrics to identify a person over a long distance without direct contact. When a person walks, it is possible to observe around 24 individual parameters and movements that form the uniqueness of their gait.
With the advancement of Computer Vision (CV) techniques, various methods for human identification through movement in video have emerged, utilizing natural biometric characteristics such as the human skeleton, silhouette, changes in movement, and abstract features.
A gait recognition system uses the shape of the human body and the way it moves in order to identify it. The software employs CV algorithms to detect a human silhouette in video footage and analyzes its movements. These data create a behavioral model of the individual.
Gait recognition algorithm
Gait recognition technology utilizes various sources or capture devices such as cameras, motion sensors, and others to acquire data. The collected data undergo several recognition steps.
The main algorithm recognizes the gait, processes the received data, detects contours, silhouettes, and segments individual human features. Subsequently, the feature extraction algorithm comes into effect — it distinguishes one gait from another.
These algorithms can vary, and their requirements can also be different. For example, some algorithms are designed to process video signals, while others use data from sensors.
Thanks to machine learning technologies, scientists can improve recognition systems based on acquired data and models. Since each gait is unique, recognition algorithms encounter new data every time they are used. The more gait variations the system encounters, the better it analyzes future data.
For instance, if the algorithm processes two very similar gaits, the pattern recognition and silhouette segmentation algorithms are trained to discern subtle details and incorporate them into the database. This enables more accurate gait classification and improves results over time.
How does gait recognition work
Researchers compiled a database of individuals' gaits, including approximately 20,000 foot movement recordings from 127 people. These recordings were captured using special floor sensors and high-resolution cameras.
All this data was loaded into a neural network for image processing. After training, the network achieved nearly 100% accuracy in recognizing people by their gait. The ML system operated on the principle of deep residual learning, enabling the identification of a person based on the spatial and temporal characteristics of their footprints.
The most common gait recognition system is based on 4 components:
- capturing gait data
- silhouette segmentation
- contour detection
- feature extraction and classification
Capturing gait data
Gait can be captured using video cameras or wearable sensors. One prominent example of such sensors are special costumes worn by actors on set, allowing motion artists to animate characters based on their movements.
Another method of capturing gait involves the use of radar to remotely identify moving objects. Radio waves are directed towards the object of interest, reflecting off its body. The system then recognizes these reflected waves and utilizes the data for identification purposes.
Silhouette segmentation
This stage is particularly suitable for studies that use video camera recordings. A binary image of a person's silhouette is extracted from the recording and analyzed by vision-based algorithms. Silhouette segmentation simplifies the algorithm's task of processing and mapping a comprehensive image.
Contour detection
Next, the system delineates the boundaries of the human body by highlighting its contours. The methods employed to achieve this goal may vary depending on the hardware used for capturing the gait, whether cameras or wearable sensors.
Feature extraction and classification
In the final step, individual features of the gait are determined. A classifier is used to identify the person, whose data is then entered into a database and used for detection.
What do gait recognition systems analyze
Gait biometric systems capture step patterns using video images and then convert the collated data into a mathematical equation. Gait as a biometric measure can be influenced by several factors, including footwear, terrain, fatigue, and injury.
The primary determinant of gait is the size of the human skeleton. Terrain also plays a role — it can cause changes in a person's speed. Injuries and footwear are significant as well. If a person walks barefoot, their gait changes. If a person has been injured, the effect on the gait can be unpredictable. In addition, the system needs to analyze the development of a person's muscles and how tired they are.
How gait recognition is used
Gait recognition systems are mainly employed in video surveillance and come in handy in crowded areas for security purposes. These systems can quickly spot a person who is wanted by authorities, thereby helping prevent terrorist attacks and other forms of crime.
A new gait recognition technology was introduced by the Chinese company Watrix. The developers claim that the system identifies people in videos based on silhouette and gait from a distance of 50 meters, and an accuracy rate of this technology is up to 94%.
However, gait recognition extends beyond security applications. For example, this technology can become an element of the smart home ecosystem. In nursing homes, it can alert staff if a patient falls. In hospitals, the system aids in diagnosing neurological disorders and planning rehabilitation therapy, while athletes benefot from its use in training.
How accurate are gait recognition systems
Scientists at the University of Manchester in the United Kingdom have developed a high-precision human gait recognition system. Researchers believe it could potentially replace retinal scanners or fingerprints in the future. According to study leader Omar Costilla Reyes, when a person walks, the system tracks approximately 24 different parameters and movements that contribute to the uniqueness of their gait.
To create a system known as SfootBD, scientists compiled a database of twenty thousand signals received from 120 individuals while they walked. Each person's gait was measured using special pressure-sensing floor panels and a high-resolution camera. Scientists then analyzed factors such as weight distribution, walking speed, and various three-dimensional measures of each gait.
The researchers tested the system in three scenarios: airport checkpoints, workplace, and homes. The system can recognize a person it already knows with nearly 100% accuracy, with an error rate of just 0.7%. Moreover, the system effectively copes with cheating attempts — when someone tries to imitate a gait.
The SfootBD system demonstrates a significant advancement, being almost 380 times more accurate than previous developments. Earlier systems required subjects to walk barefoot on special touch panels or relied on 3D-shooting technology to compare individuals' walking patterns with surveillance camera footage.
Advantages of gait recognition
Gait recognition technology is less “touchy-feely” than other biometric verification systems, such as retinal scans or fingerprints. Thus, it is non-invasive and can be applied without user consent. Moreover, the success rate of this technology is high — the error rate is only 0.7%.
Disadvantages of gait recognition
- The new technology has its limitations: it requires special sensing panels and a high-resolution camera. Additionally, the system can only recognize individuals whose data have been pre-recorded and stored in its database, limiting its current applications.
- There are also concerns about the potential for covert use without individuals' knowledge, raising significant issues related to security and privacy.
- Moreover, since the technology is still in its developmental stage, its results may not always be completely reliable. For example, many internal factors such as diseases or psychological conditions can have a more pronounced impact on gait than external factors, potentially affecting recognition accuracy.
Summary
Of course, the new system is not universal. Firstly, to capture and record a person's gait into the database, they must be placed in a room equipped with floor sensors and monitored by high-resolution cameras. Secondly, the algorithms can only recognize individuals already in the database, making scalability of the technology quite challenging.
On the other hand, the technology requires no direct contact with subjects and is easily adaptable for use in public places. Furthermore, the new algorithm boasts a negligible error rate, which means that, with proper development, such a system could be very promising.
FAQ
Now let us take a look at some of the most popular questions regarding gait recognition technology.
How do gait biometrics work?
Neural networks can identify patterns in a person's gait, enabling recognition and identification with nearly 100% accuracy.
Gait is digitized using high-resolution sensors and cameras. The system analyzes all collected data, including weight distribution, walking speed, and three-dimensional features of each gait style.
The result is a fully-fledged «cast» of the person and their walking pattern, which is sored in a database for future use.
Is gait recognition accurate?
The experiments revealed that this technology recognizes a person with nearly 100% accuracy, with an error rate of just 0.7%. Therefore, we can conclude that gait recognition technology achieves a high level of accuracy. Nevertheless, since the technology is still in development, it may be premature to draw any definitive conclusions.
What features are observed in gait recognition?
Gait recognition systems capture step patterns using video images. Several factors are critical for accurate recognition, including footwear, terrain, fatigue, injuries, and the development of a person's muscles.
How can gait recognition be fooled?
Given that the technology is still in its infancy, it may still be possible to deceive it, but it would require significant effort. The system effectively handles attempts to mimic another person's gait. After all, each person's gait is unique, making them more challenging to replicate than one might assume.
