Learning Resolution-Invariant Deep Representations for Person Re-Identification

Abstract

Most existing person re-identification (ReID) methods have good feature representations to distinguish pedestrians with deep convolutional neural network (CNN) and metric learning methods. However, these works concentrate on the similarity between encoder output and ground-truth, ignoring the correlation between input and encoder output, which affects the performance of identifying different pedestrians. To address this limitation, We design a Deep InfoMax (DIM) network to maximize the mutual information (MI) between the input image and encoder output, which doesn't need any auxiliary labels. To evaluate the effectiveness of the DIM network, we propose end-to-end Global-DIM and Local-DIM models. Additionally, the DIM network provides a new solution for cross-dataset unsupervised ReID issue as it needs no extra labels. The experiments prove the superiority of MI theory on the ReID issue, which achieves the state-of-the-art results.

CrDoCo: Pixel-Level Domain Transfer With Cross-Domain Consistency

“…space. In the context of image classification, several methods [12,13,29,30,46,47,4] have been developed to address the domain-shift issue. For semantic segmentation tasks, existing methods often align the distributions of the feature activations at multiple levels [17,19,45].…”

DRL-GAN: Dual-Stream Representation Learning GAN for Low-Resolution Image Classification in UAV Applications

“…Resolution-invariant representations are crucial for crossresolution recognition [21], [22], [23], [24], [25]. Image representations, which are invariant to the image resolution change, were proposed and used to compensate the missing details in LR images for boosting the performance of the person re-ID task in [23].…”

Multi Scale-Adaptive Super-Resolution Person Re-Identification Using GAN

“…Given their promising results, it relied on the annotation of the foreground mask to direct the learning of image recovery for each training image. Chen et al [23] proposed resolution adaptation and reidentification Network (RAIN). They used adversarial loss and reconstruction loss to reduce the difference between different resolution deep features by aligning the feature distributions of HR and LR images.…”