Device and method for cross-domain person re-identification based on domain-invariant features
A pedestrian re-identification and domain-invariant technology, applied in the field of pedestrian re-identification, can solve the problems of time-consuming and laborious data collection in the target domain, and the inability to collect data in the target domain, etc., to enhance cross-domain generalization capabilities and improve cross-domain generalization Ability, the effect of cross-domain person re-identification performance improvement
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0055] The present invention designs a cross-domain pedestrian re-identification device based on domain-invariant features. The cross-domain pedestrian re-identification device is arranged after the residual module of the domain-invariant feature extraction network, such as figure 2 As shown, it includes a restored feature module for obtaining restored features, a feature enhancement module for obtaining discriminative features, and a feature stacker for stacking the restored features and discriminative features to obtain complete output features
[0056] The recovery feature module is provided with:
[0057] The instance normalization module IN is used to normalize the input original features to obtain the normalized features of the instance;
[0058] Feature Residual Calculator It is used to calculate the residual between the original feature of the input and the feature after instance normalization, and obtain the residual feature;
[0059] The first attention mechani...
Embodiment 2
[0065] This embodiment is further optimized on the basis of the above-mentioned embodiment, and the cross-domain pedestrian re-identification method based on domain-invariant features is implemented by using the cross-domain pedestrian re-identification device based on domain-invariant features, such as figure 2 shown, including the following steps:
[0066] 1) The original features of the input reduce the domain difference between the sample features through the instance normalization module IN, and obtain the features after the instance normalization;
[0067] 2) Utilize the feature residual calculator Perform residual calculation on the input original features and the features after instance normalization to obtain residual features;
[0068] 3) Using the residual feature to extract features related to pedestrian identity information adaptively based on the channel attention mechanism and the spatial attention mechanism by using the first attention mechanism module (incl...
Embodiment 3
[0075] This embodiment is further optimized on the basis of the foregoing embodiment, and the same parts as the foregoing technical solutions will not be repeated here, such as figure 2 As shown, in order to better realize the cross-domain pedestrian re-identification method based on the domain-invariant feature of the present invention, the original feature of the input is set to x, and x∈R b×c×h×w , where b, c, h, and w represent the batch size, the number of channels, the height and width of the feature map, respectively, and R b×c×h×w is a b×c×h×w-dimensional matrix, then the complete output feature after the attention and style normalization module ASN (that is, the cross-domain person re-identification device) is y∈R b×c×h×w ;
[0076] The normalized feature of the instance is set as x 1 , in the step 1), the original feature of the input adopts the following formula to obtain the feature after the instance normalization through the instance normalization module IN: ...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More