A gesture recognition system and method for the game of guessing and punching

Abstract

The invention discloses a gesture recognition system and a gesture recognition method for a guessing and rowing fist game. Being used as a gesture recognition area, a gesture placing plate is arrangedon the gesture placing plate by hand, four independent capacitive sensors are arranged under the gesture place plate, four independent capacitive sensors are connected to an input end of a sensing chip, and an output end of the sensing chip is connected to a control microprocessor. Four independent capacitive sensors are arranged in a palm hand shape, wherein three independent capacitive sensorsare respectively arranged in a middle finger, a ring finger and a little finger position area of the palm, and another independent capacitive sensor is respectively arranged in a position area composed of a thumb and an index finger of the palm. The invention realizes non-contact gesture input and outputs accurate result, has high realization efficiency, high recognition accuracy and lower cost, and is very simple and effective.

Description

technical field [0001] The present invention relates to a human body behavior recognition system and method, in particular to a gesture recognition system and method for the game of guessing and punching. Background technique [0002] Under the current machine learning technology, learners directly absorb the information provided by the environment without any reasoning or other knowledge conversion. However, machine learning also has some problems that cannot be ignored. The first is the collection and selection of data sets. The larger the data set, the less dependent on the environment, but the higher the hardware performance and technical requirements. In the field of gesture recognition, because each person's palm size is different, if you want to collect data from the general population, the data set and hardware requirements may be too high. Therefore, the adjustability of machine learning to gesture recognition is weak, and the implementation method is too complica...

AI Technical Summary

Problems solved by technology

In the field of gesture recognition, because each person's palm size is different, if you want to collect data from the general population, the requirements for data sets and hardware may be too high

Therefore, the adjustability of machine learning to gesture recognition is weak, and the implementation method is too complicated, which seems a bit overkill

[0003] Traditional image-based gesture recognition needs to be equipped with a camera, and is easily disturbed by environmental factors such as light. At the same time, most of them use neural network algorithms, which have high requirements for the computing power of the processor, and have the disadvantages of high cost and being easily affected by light.

[0004] Based on the above two points, machine learning and neural networks are used to complete this task for the boxing game. In the early stage, a large number of data sets from different groups of people need to be collected, and hardware devices with powerful computing capabilities are needed as support, and the result of gesture recognition may be There will be a large error as the environmental conditions change, and the investment required is too large

In addition, if traditional image recognition technology is used, such as feature point detection, to achieve high recognition accuracy, the algorithm may be very complicated and the computing power requirements of the hardware will be further increased, resulting in recognition delay will also increase

Images