Gear recognition

from Wikipedia, the free encyclopedia

The transition detection means a biometric method , the different characteristics of human gait analyzed and using a comparison to other records people clearly identified. Different detection systems are based on several types of sensor technology .

Analytical methods

As with any biometric method, in order to verify a data set, it is compared with another data set using a suitable algorithm. If the degree of similarity between the two is above a predetermined value, the system classifies the user as verified and grants access. Alternatively, a user can also be compared with every user already stored in the system in order to identify them.

Machine vision

With the help of optical sensors, a picture of a person to be analyzed is created and relevant data of the picture are extracted using image processing algorithms . A static background, for example a wall, is recommended. Due to the high computing power involved, some approaches are limited to the human silhouette . In most cases, the recording is also divided into several sequences, each representing a step. The distance between the sensor and human being, which is comparatively large for biometric recognition methods, is generally advantageous; the disadvantage is the high susceptibility to interference of the recording (lighting conditions, distance between human and sensor).

Floor sensors

With the help of sensors attached under the floor plates, the force is measured which the foot transmits to the floor plate when stepping on it. Thus, for example, a pressure profile of the person appearing can be created in a process, which can be used to identify the person. Measured values ​​used in other procedures include body weight and foot shape.

Wearable sensors

In this approach, portable accelerometers and / or gyrometers are placed in different parts of the body of the person to be recognized, including the feet, belt or trouser pockets. Most of the sensors used are already built into modern smartphones and are therefore accessible to the general public. In this context, for example, passive authentication on the smartphone would be possible, which is an alternative to entering a password.

Disruptive factors

The human gait is unique for every person, but it is practically always falsified by external influences. These include the clothing worn with a focus on the footwear, various physiological parameters of the runner such as short-term health complaints, increasing or decreasing body weight or food consumed as well as varying running pace or objects carried. Correspondingly, aisle recognition is not suitable as the sole option for authentication on a system with restricted access, but should always be used in conjunction with alternative authentication options .

Individual evidence

  1. ^ Liu, Zongyi & Sarkar, Sudeep : Simplest representation yet for gait recognition: Averaged silhouette. (PDF, 286 kB) 2004, accessed on February 2, 2015 (English).
  2. Orr, Robert J. & Abowd, Gregory D .: The Smart Floor: A Mechanism for Natural User Identification and Tracking. (PDF, 121 kB) (No longer available online.) 2000, archived from the original on February 2, 2015 ; accessed on February 2, 2015 . Info: The archive link was inserted automatically and has not yet been checked. Please check the original and archive link according to the instructions and then remove this notice. @1@ 2Template: Webachiv / IABot /