Person reidentification (ReID) requires the discriminative features of an entire pedestrian image to handle the problems of inaccurate person bounding box detection, background confusion, and occlusion. Many recent person ReID methods have attempted to learn the feature information of an entire pedestrian image through parts feature representations, but it is often too time consuming. Person ReID relies on discriminative pedestrian features, and different spatial scales can distinguish features by differing degrees. We propose an innovative and effective adaptive spatial scale person ReID network model based on the residual neural network (ResNet) of an instance batch normalization. Through experimental visualizations, pedestrian features extracted by ResNet from four layers are analyzed, and two layers with discriminative features are selected. Using an adaptive dimension adjustment module, different spatial scale features are aggregated and merged by the aggregation layer. To effectively learn spatial channel correlations and prevent overfitting, a multilayer distribution normalization processing module is designed to implement end-to-end training and evaluate the person ReID networks. Compared with other methods, this network model showed excellent performance on four public person ReID datasets and is superior to most current methods.
As Hadamard measurement matrix cannot be used for compressing signals with dimension of a non-integral power-of-2, this paper proposes a construction method of block Hadamard measurement matrix with arbitrary dimension. According to the dimension N of signals to be measured, firstly, construct a set of Hadamard sub matrixes with different dimensions and make the sum of these dimensions equals to N. Then, arrange the Hadamard sub matrixes in a certain order to form a block diagonal matrix. Finally, take the former M rows of the block diagonal matrix as the measurement matrix. The proposed measurement matrix which retains the orthogonality of Hadamard matrix and sparsity of block diagonal matrix has highly sparse structure, simple hardware implements and general applicability. Simulation results show that the performance of our measurement matrix is better than Gaussian matrix, Logistic chaotic matrix, and Toeplitz matrix.
In order to improve the accuracy of global motion estimation (GME), a new method for GME combining with motion segmentation is proposed in this paper. The proposed method removes motion vector (MV) outliers and implements initial motion segmentation by analyzing properties of motion vectors. Using the filtered MV field, global motion parameters were estimated, and then the difference frame was generated by global motion compensation(GMC ). According to the movement difference between the background and the foreground regions and movement consistency in the same region, the absolute sum of difference frame in every block was calculated, and thus adaptively generating the threshold value to detect motion regions. MVs in the motion regions were rejected as outliers for GME, and iterative computations between GME and motion segmentation were performed successively. Experimental results demonstrate that the proposed approach can effectively extract motion regions, thus enhancing the accuracy of GME.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.