Sensing method for gesture and spatial location of hand

Abstract

The invention provides a human-machine interaction technology for the hand gesture recognition based on variable number of cameras by adopting an infrared light source. In the method, a user does not need to wear any assistive device, and the method has the outstanding characteristics of wide application range, comprehensive motion sensing capabilities, low production cost, low computation complexity and the like. In the invention, the gesture and location of a hand are modeled by using a virtual reality method, and a template database is generated. In practical use, background-removed input images of a plurality of cameras are compared with items in the template database, and an item with the minimum contrast difference in the database is regarded as an initial recognition result of the gesture of the hand. In order to obtain a more stable recognition result, a smoothing filter method is adopted to correct the initial recognition result so as to provide more stable recognition data for the user. In addition, sequences of hand motions in a continuous period can be recognized by the human-machine interaction technology, and rich options are provided for the human-machine interaction.

Description

technical field [0001] The present invention generally relates to improved hand gesture and position sensing methods. More specifically, the present invention refers to a multi-camera system-based method for sensing hand posture and position using infrared light, which is used to provide improved human-computer interaction technology. Background technique [0002] Over the years, with the popularization and development of multimedia technology, people have been unremittingly exploring new human-computer interaction technology. The use of limbs, gestures and other intuitive ways to complete computer manipulation has become a technology hotspot. The human hand is a complex executive mechanism with high flexibility, rich expressiveness, and fine manipulation. However, these characteristics also make the recognition and tracking of gestures a major challenge in computer research. [0003] The recognition of hand movements can be achieved in a variety of ways. US20100199228A1 ...

AI Technical Summary

Problems solved by technology

This method is relatively intuitive, but requires a complex mathematical calculation process, and cannot identify and track the opponent's spatial position

[0005] In addition, in the field of hand motion perception, there are still several major technical challenges: First, how to effectively extract the hand area from the acquired image

Some technical solutions perform segmentation processing by identifying human skin color, but they are greatly affected by ambient light conditions, and special processing is required for skin colors of different races

In other solutions, some colored logos or colored gloves are worn on the hands, but they are also affected by certain ambient light and reduce the comfort and convenience of using the equipment.

Secondly, the hand image in the image has a serious self-occlusion phenomenon, that is to say, in an image, fingers and palms may cover each other, making it difficult for the computer to judge the specific gesture of the finger based on the information in the image

Images