Calligraphy attitude automatic identification method based on Wi-Fi signals

Abstract

The invention discloses a calligraphy attitude automatic identification method based on Wi-Fi signals. The method comprises the following steps: calligraphy character attitude data is acquired through Wi-Fi signals; feature extraction and stroke segmentation are performed on the data in turn; the data is reconstructed using a learning algorithm; and the writing font is identified by comparing the data with a standard database. The method is implemented using the existing commercial equipment. There is no need to modify the hardware. There is no need for extra deployment for users. Wireless signals are easy to obtain, and can be emitted by opening a hotspot via an ordinary smart mobile device. The CSI value of the physical layer is analyzed using the existing wireless local area network in a non-intrusive and device-independent way. There is no need to change the wireless signal communication protocol. The method has a wide application prospect.

Description

technical field [0001] The invention relates to the technical field of wireless signal tracking perception and human-computer interaction, in particular to a method for automatic recognition of calligraphy gestures based on Wi-Fi signals. Background technique [0002] Chinese calligraphy is the art of writing Chinese characters with a special Chinese conical brush, and is an important carrier of Chinese culture. In 2015, the Ministry of Education issued the "Guidelines for Improving the Education of Excellent Chinese Traditional Culture", which further clearly requires students to be able to write Chinese characters in a standardized way, to copy the calligraphy of famous masters, and to experience the beauty and artistic conception of calligraphy. For a long time, the traditional on-the-spot observation-style calligraphy teaching method requires teachers to guide students to make posts, demonstrate and correct calligraphy strokes one by one. In the process of practice, the...

AI Technical Summary

Problems solved by technology

[0004] In the existing technology, infrared technology can only identify activities in a specific area, which has high requirements on infrastructure and expensive equipment; based on vision methods, users need to be photographed and videotaped, which will leak some user privacy and require a large amount of calculation. It is suitable for the recognition of the line-of-sight range, and it is prone to dead angles and is affected by light, obstacles, etc. Although the use of LEDs and light sensors can achieve millimeter-level positioning accuracy, the visual angle is limited, and lighting conditions also have a great impact

Attitude recognition based on sound signals using Doppler frequency shift has no tracking ability and can only recognize predefined attitudes; among them, the 60GHz radio frequency signal can achieve millimeter-level tracking accuracy, but it requires expensive professional equipment and is not universal. Ultrasound is easy to attenuate during propagation, the range of recognition is limited, and the speed must be greater than a certain value to detect Doppler frequency shift, and additional hardware is required; using voice recognition technology to interact with devices, although it has been widely used in smart homes and other aspects Some applications, but not yet widely promoted

Behavior recognition based on dedicated sensors can identify fine-grained behaviors, but it is inconvenient to install and carry, and it is expensive, and it is not suitable for Device Free scenarios

The traditional Wi-Fi signal is used for positioning and attitude recognition, mainly using the RSSI value of the Wi-Fi signal, but the RSSI value can only be used to identify large-scale gestures, because fine-grained gesture changes have little impact on the RSSI value. It is hardly observable, so the use of RSSI values ​​​​is invalid for fine-grained gesture recognition

Images