Pedestrian re-identification method based on fusion of attitude estimation and viewpoint mechanism
A pedestrian re-identification and attitude estimation technology, applied in the field of pedestrian re-identification, can solve the problem of occlusion of key areas of pedestrians, achieve the effects of overcoming changes in pedestrian posture, alleviating alignment difficulties, and improving efficiency and accuracy
Pending Publication Date: 2019-07-12
NANJING UNIV OF TECH
View PDF6 Cites 15 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
It is used to solve various criminal investigation and missing person finding problems. In order to achieve this purpose, the method considers the pedestrian's posture structure information, adopts the pose estimation network Openpose to extract pedestrian key points, and considers the viewpoint position of the camera. According to the image's The shooting angle generates the key area of the pedestrian's body, which helps to align the body area features in the image and avoids the problem that the key area of the pedestrian is blocked when the camera captures the side of the pedestrian
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
[0008] see figure 1 Shown:
[0009] 1. A pedestrian re-identification method based on pose estimation and viewpoint mechanism fusion, characterized in that the method comprises the following steps:
[0010] Step 1: Collect pedestrian re-identification datasets, obtain public datasets for pedestrian re-identification, distinguish between training sets and test sets, and proceed to step 2;
[0011] Step 2: Pedestrian re-identification data set training, the training set of the pedestrian re-identification public dataset is used as the training set of the deep convolutional neural network model, and finally the pedestrian re-identification model is obtained to enter step 3;
[0012] Step 3: Feature extraction and fusion. According to the pedestrian re-identification model obtained in step 2, the global features and local features in the network are extracted from the query image and the library image. After an image is input into the network, the feature is embedded in the netwo...
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 designs a pedestrian re-identification method based on fusion of attitude estimation and viewpoint mechanism, which is used for solving the problem of criminal investigation and missingperson looking. In order to solve the problem, posture structure information of pedestrians is considered; a pedestrian key point is extracted by adopting a posture estimation network Openpose, meanwhile, the viewpoint position shot by a camera is considered, a pedestrian body key area is generated according to the shooting angle of an image, body area characteristics in the image are helped to bealigned, and the problem that the pedestrian key area is blocked when the camera shoots the side face of a pedestrian can be avoided. Finally, the local feature map and the global feature map are fused through a feature fusion sub-network to merge regional features of different semantic levels. A corresponding model is obtained by training a pedestrian data set, and all pedestrians matched with atarget pedestrian are predicted by loading the model and adopting a neural network structure.
Description
technical field [0001] The invention relates to the field of computer vision and image retrieval, and specifically designs a pedestrian re-identification method based on the fusion of attitude estimation and viewpoint mechanism. Background technique [0002] Person re-identification (Person re-identification, Re-ID) originated from multi-camera tracking, and is used to determine whether pedestrians in different images captured in non-overlapping fields of view belong to the same person. Pedestrian re-identification involves computer vision, machine learning, pattern recognition and other disciplines, and can be widely used in intelligent video surveillance, security, criminal investigation and other fields. In recent years, pedestrian re-identification technology has attracted extensive attention from academia and industry, and has become a research hotspot in the field of computer vision. Due to the characteristics of both rigid and flexible objects, the appearance of pedes...
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 Applications(China)
IPC IPC(8): G06K9/00G06K9/62
CPCG06V40/103G06F18/253G06F18/214
Inventor 徐曾春裴嘉震胡平
Owner NANJING UNIV OF TECH