A Dynamic Gesture Recognition Method Based on Hidden Markov Model Incremental Learning
A dynamic gesture and recognition method technology, applied in character and pattern recognition, instruments, computer components, etc., can solve problems such as spending a lot of time and energy, adjusting model parameters, etc., to achieve improved recognition accuracy, high recognition rate, and good quality The effect of recognition effect
Active Publication Date: 2016-09-14
NANJING UNIV
View PDF3 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0011] (2) Once the newly identified samples complete the identification task, they will no longer have other functions, and the system cannot adjust the model parameters in real time according to the newly added samples to make it more adaptable to new scenarios
At this time, it will take a lot of time and energy to retrain all samples (including existing training samples and new recognition samples) to adapt to the new situation.
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
[0137] In the embodiment, the gesture operation trajectory of 10 Arabic numerals from 0 to 9 is recognized. The operator simulates the trajectories of 10 Arabic numeral strokes in the spatial area with the human hand in front of the camera. These quantized gesture trajectories are used for model training, gesture recognition and incremental learning.
[0138] In the training phase, each Arabic numeral is trained using 40 video streams, that is, 40 training samples, so that the total number of training samples is 400.
[0139] In the recognition stage, the number of recognition samples for each Arabic numeral ranges from 70 to 100 (note: pose "1" was used as a test video during the experiment, so a large number of sample libraries were recorded), and the training samples and recognition The number of samples is shown in Table 1:
[0140] Table 1 List of training samples and recognition samples in the experiment
[0141]
0
1
2
3
4
5
6
7
8
...
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
The invention discloses a dynamic gesture recognition method based on hidden Markov model self-incremental learning, comprising the following steps: (1) gesture detection and tracking; (2) feature extraction and vector quantization; (3) model training and gesture Recognition; (4) Incremental learning. The present invention can accurately recognize the dynamic gesture operation completed by the gesture operator in front of the camera by using the dynamic gesture recognition method based on hidden Markov model incremental learning proposed by the present invention, and can use the recognized gesture data for old Incremental learning of the model to adjust the model parameters, so that the old model can dynamically adapt to the new changes in the future gesture data, and have better adaptability to the adjustment and change of the gesture data, so that the model can continue to follow The adjustment of gesture data has better robustness to future unknown gesture recognition.
Description
technical field [0001] The invention relates to the fields of computer vision, image processing, pattern recognition and the like, in particular to a dynamic gesture recognition method based on hidden Markov model self-incremental learning. Background technique [0002] With the rapid progress and development of science and technology, computer science also takes off rapidly. At present, while the computer field is developing toward higher speed, higher efficiency, and higher computing rate, it is also striding forward toward the field of human-computer interaction that is more convenient, simpler, and more comfortable. [0003] Especially with the hot sales of a series of electronic consumer products such as mobile phones and tablet computers, providing better man-machine interface and facilitating more natural and harmonious communication between people and computers has become a quite potential economic detonation in the computer field. point. [0004] At present, in th...
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): G06K9/66
Inventor 申富饶胡孟赵金熙
Owner NANJING UNIV