When deploying re-identification (ReID) models in the field of public safety, understanding the robustness of models to various types of corrupted images is crucial. Unfortunately, in the real world, images are always contaminated (e.g., noise, blur, and weather changes), which is ignored by existing visible-infrared person re-identification (VI-ReID) models. The performance of existing models tested in corrupted scenes is severely degraded. Therefore, learning corruption-invariant representations for corrupted images in VI-ReID is valuable and deserves further investigation. We design a polymorphic masked wavelet graph convolutional network for VI-ReID under corrupted scenes. Firstly, a cross-modality data augmentation algorithm is designed to construct a mixed image set that merges multi-modality attributes to improve robustness against interference. Secondly, a dual-branch network consisting of a global branch and a graph structure branch is designed. The global branch extracts overall information. While the graph structure branch is a wavelet-based graph convolutional module that utilizes the robustness of human structural information to corruptions and modalities, it can filter noise and extract discriminative features specifically targeted for cross-modality scenes. Finally, the global branch and the graph structure branch are integrated, and modality consistency loss is designed to match the branches with hetero-center triplet loss. Experiments show that our method can effectively alleviate degradation problems under corrupted environments such as noise, blur, digitization, and weather changes, and achieve state-of-the-art on corrupted datasets. Besides, it still maintains good performance on clean datasets, facilitating the reliable deployment of VI-ReID in real-world scenarios.

The transmission of the orbital angular momentum (OAM) beam has attracted a lot of attention in the field of free-space optical (FSO) communication. Usually, after transmitting in atmospheric turbulences, the helical phase-front of OAM beams will be severely distorted, and there will exist the intermode crosstalk. As a result, the performance of the communication system will degrade significantly. In this paper, we have investigated the influences of the level of the turbulence strength to the transmitting OAM beams by changing the refractive-index structural parameter of $C_n^2$ and the number of turbulence random phase screens of N in simulation environment. Then, by adopting the Gerchberg-Saxton (GS) algorithm, which can be used to compute the pre-compensation phase and correct the distorted OAM beams, the retrieving performances of transmitting single and multiplexed OAM beams under different turbulence strengths were also investigated. The simulation results show that with increasing the atmospheric-turbulence strength levels determined by the parameters $C_n^2$ and N, the retrieving performances decrease dramatically. When the turbulence strength level is selected within an appropriate range, the OAM beams can be effectively retrieved by adopting GS algorithm and observing the power density spectrum. Notably, the retrieving performance for the transmission of a single OAM beam is better than that of the multiplexing OAM beam.

With the prevalence of dual-mode cameras in surveillance systems, visible-infrared person re-identification (VI-ReID) has become an emerging topic. Existing studies of VI-ReID roughly fall into three categories: straightforwardly extracting features, improving loss functions, and conducting visible-infrared modality generation. The generation methods avoid the shortcoming of the former two that training models are generally vulnerable to parameter changes. However, these generation methods are usually based on spatial domain and are unavoidable to damage the original information of images. To tackle these limitations, we propose a novel frequency-domain simulated multispectral (FSMS) modality and visible-FSMS-infrared collaborative learning. FSMS modality consists of three-channel images generated by a channel-level reconstruction of visible images, primarily based on the nonsubsampled contourlet transform (NSCT) cooperating with a lightweight network. The generation exploits crucial spectral information and edge information contained in frequency domain. Then, we design a multi-modality network to conduct the tri-modality collaborative learning where FSMS modality is utilized as an intermediate, thereby preserving the original spatial structure of images. Additionally, a dynamic-weight tri-modality heterogeneous retrieval (THR) loss and a modality-shared classification (MSI) loss are devised to mine discriminative modality-invariant features. A cross-modality invariant (CMI) constraint for further exploring triplet-wise relationships and an intra-modality regularizer for relatively stable convergence are introduced. Finally, experimental results show that our algorithm significantly outperforms the latest state-of-the-arts by 5.7% and 4.4% CMC-1 accuracy on two mainstream benchmark datasets, respectively. And the reasons underlying the observed increase in performance are deeply discussed.

Based on the Monte Carlo (MC) algorithm, we simulate the evolutions of different types of the polarized lights in the broad-band range from visible to infrared in foggy environments. Here, we have constructed two scattering systems to simulate the transmission characteristics of the polarized lights: (1) A monodisperse system based on five types of particles with the sizes of 0.5, 1.0, 2.5, 4, and 5 µm, respectively; (2) a polydisperse system based on scattering particles with a mean value (size) of 2.0 μm. Our simulation results show that linearly polarized light (LPL) and circularly polarized light (CPL) exhibit different advantages in different wavelengths and different scattering systems. The polarization maintenances (PM) of the degree of circular polarizations (DoCPs) are better than those of the degree of linear polarizations (DoLPs) for most incident wavelengths. CPL is not superior to LPL in the strong-absorption wavelengths of 3.0µm, 6.0µm, and long infrared. Here, when the wavelength is closer to the particle sizes in a system, the influence on propagating polarizations will be more obvious. However, the difference in the degree of polarization (DoP) between the resulting CPL and LPL is positive at these points, which means the penetrating ability of CPL is superior to that of LPL in these scattering systems. We have also simulated the extinction efficiency Q_ext and the scattering index ratio Q_ratio as functions of both wavelength and particle size for analyzing polarization’s transmission characteristics. Our work paves the way of selecting the optimal incident wavelengths and polarizations for concrete scattering systems.

The main challenge of person re-identification (re-id) lies in the strikingly discrepancy between different camera views, including illumination, background and human pose. Existing person re-id methods rely mostly on implicit solutions, such as seeking robust features or designing discriminative distance metrics. Compared to these methods, human solutions are more straightforward. That is, imagine the appearance of the target person under different camera views before matching target person. The key idea is that human can intuitively implement viewpoint transfer, noting the association of the target person under different camera views but the machine failed. In this paper, we attempt to imitate such human behavior that transfer person image to certain camera views before matching. In practice, we propose a conditional transfer network (cTransNet) that conditionally implement viewpoint transfer, which transfers image to the viewpoint with the biggest domain gap through a variant of Generative Adversarial Networks (GANs). After that, we obtain hybrid person representation by fusing the feature of original image with the transferred image then perform similarity ranking according to cosine distance. Compared with former methods, we propose a human-like approach and obtains consistent improvement of the rank-1 precision over the baseline in Market-1501, DukeMTMC-ReID and MSMT17 dataset by 3% ,4%,4%, respectively.