Person Re-Identification by Dual-Regularized KISS Metric Learning

PROJECT TITLE :

Person Re-Identification by Dual-Regularized KISS Metric Learning

ABSTRACT:

Person re-identification aims to match the photographs of pedestrians across completely different camera views from different locations. This is often a difficult intelligent video surveillance downside that remains an active area of research thanks to the need for performance improvement. Person re-identification involves two main steps: feature illustration and metric learning. Although the keep it simple and simple (KISS) metric learning methodology for discriminative distance metric learning has been shown to be effective for the person re-identification, the estimation of the inverse of a covariance matrix is unstable and indeed might not exist when the training set is little, resulting in poor performance. Here, we have a tendency to gift twin-regularized KISS (DR-KISS) metric learning. By regularizing the 2 covariance matrices, DR-KISS improves on KISS by reducing overestimation of enormous eigenvalues of the 2 estimated covariance matrices and, in doing so, guarantees that the covariance matrix is irreversible. Furthermore, we provide theoretical analyses for supporting the motivations. Specifically, we tend to 1st prove why the regularization is important. Then, we prove that the proposed technique is strong for generalization. We have a tendency to conduct intensive experiments on three challenging person re-identification datasets, VIPeR, GRID, and CUHK 01, and show that DR-KISS achieves new state-of-the-art performance.

Did you like this research project?

To get this research project Guidelines, Training and Code... Click Here