A target gesture recognition method, device and electronic device
A technology of target posture and recognition method, applied in the field of image processing, can solve the problems of low accuracy of posture recognition, less information, limited feature points, etc.
Active Publication Date: 2021-04-09
HANGZHOU HIKVISION DIGITAL TECH
View PDF9 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0003] However, the feature points in the feature point model are limited, and the feature point model carries less information, resulting in lower accuracy of pose recognition
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 approach
[0211] As an implementation, the device may also include:
[0212] The second input module (not shown in the figure) is used to input the image features and the shape features and / or structural features of the target into the environmental neural network model obtained in advance to obtain the environmental factors of the environment where the target is located ; Wherein, the environmental neural network model is: according to the environmental factors corresponding to the preset feature samples, the network parameters of the environmental neural network with the preset structure are trained, and the feature samples include image features and shape features of the target and / or structural features.
[0213] As an implementation, the device may also include:
[0214] A fusion module (not shown in the figure), configured to fuse the shape features and / or structural features of the target with the image features to obtain fused features;
[0215] The first input module 804 can ...
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 embodiment of the present invention provides a target posture recognition method and device, pre-training to obtain the neural network model: according to the posture information corresponding to the preset feature samples, the network parameters of the neural network with a preset structure are trained to obtain the neural network model; When performing target gesture recognition, determine the image features of the image area where the target is located, and the shape features and / or structural features of the target, input the image features and the shape features and / or structural features into the neural network model, and output the target's gesture; It can be seen that in this solution, the target posture is recognized based on the neural network model. First, the neural network model is trained based on a large number of samples, and the information carried is richer than that carried by the feature point model, and the recognition accuracy is higher. Second, The network parameters of the neural network model in this solution carry image features, and information about the shape features and / or structural features of the target, which further improves the accuracy of recognition.
Description
technical field [0001] The present invention relates to the technical field of image processing, in particular to a target posture recognition method, device and electronic equipment. Background technique [0002] In the process of image processing, it is usually necessary to recognize the pose of the target in the image, such as face angle recognition, human pose recognition, vehicle pose recognition, etc. Existing recognition schemes usually include: for the same type of target (such as human face, human body, vehicle, etc.), pre-constructing the feature point model under each posture; The feature points are matched with the constructed feature point models, and the pose corresponding to the successfully matched model is the pose of the target. [0003] However, the feature points in the feature point model are limited, and the feature point model carries less information, resulting in lower accuracy of pose recognition. Contents of the invention [0004] The purpose o...
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): G06K9/62
CPCG06T7/73G06T2207/20084G06T2207/20081G06F18/22G06F18/214
Inventor 郑钢钮毅
Owner HANGZHOU HIKVISION DIGITAL TECH