Human action recognition method based on convolutional neural network
A convolutional neural network and human action recognition technology, applied in the field of human action recognition based on convolutional neural network, can solve the problems of limiting the upper limit of the recognition rate, easy to cause confusion, etc., to achieve the effect of improving accuracy and robustness
Inactive Publication Date: 2021-07-23
CIVIL AVIATION UNIV OF CHINA
View PDF3 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, after the depth information of the action is mapped to the two-dimensional space representation, the action is easily confused during the classification process, which limits the upper limit of the recognition rate of this type of method.
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0017] The present invention will be described in further detail below in conjunction with the examples.
[0018] Such as figure 1 As shown, the human action recognition method based on the convolutional neural network provided by the present invention includes the following steps carried out in order:
[0019] (1) Select part of the depth images in the data set as training samples, and the rest of the depth images as test samples, and then use the spatial structure dynamic depth image technology to map the four-dimensional information of the depth images in the data set to two-dimensional space to obtain two-dimensional images, It is used for subsequent classification, thus transforming the human action recognition problem into an image classification problem;
[0020] Using SSDDI technology can convert each depth image into 6 different 2D images, divide these 6 2D images into 3 groups, the groups are trunk, limbs and joints, each group consists of two 2D images, namely DDIF...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
A method for human action recognition based on convolutional neural network. It includes selecting part of the depth images in the data set as training samples, and the rest of the depth images as test samples, and using spatial structure dynamic depth image technology to map the four-dimensional information of the depth images in the data set to two-dimensional space to obtain two-dimensional images; Convolutional neural network; use the two-dimensional image in the training sample to train the convolutional neural network; input the two-dimensional image in the test sample into the above-mentioned trained convolutional neural network to obtain three sets of output vectors, and then perform intra-group Fusion, then fusion between groups, and finally complete the steps of human motion recognition. The method of the invention can be used as the basis of pattern recognition and artificial intelligence, and has great significance for human body action recognition.
Description
technical field [0001] The invention belongs to the technical field of human motion recognition, and in particular relates to a human motion recognition method based on a convolutional neural network. Background technique [0002] At present, human action recognition has been widely used in intelligent monitoring, human-computer interaction, video retrieval, virtual reality, etc., so it has always been an active research direction in the field of computer vision. In previous studies, many research methods on human action recognition have focused on traditional RGB color videos. In recent years, the release of Microsoft Kinect has brought new opportunities to this field. The Kinect device can collect depth images in real time. Compared with traditional color images, depth images have many advantages. For example, depth image sequences are essentially four-dimensional space , can contain richer motion information, is insensitive to changes in lighting conditions, and can esti...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/00G06K9/62G06N3/04
CPCG06V40/20G06N3/045G06F18/214
Inventor 张良李玉鹏刘婷婷
Owner CIVIL AVIATION UNIV OF CHINA