Nanjing University

About: Nanjing University is a education organization based out in Nanjing, China. It is known for research contribution in the topics: Catalysis & Population. The organization has 85961 authors who have published 105504 publications receiving 2289036 citations. The organization is also known as: NJU & Nanking University.

Selective Convolutional Descriptor Aggregation for Fine-Grained Image Retrieval

Abstract: Deep convolutional neural network models pre-trained for the ImageNet classification task have been successfully adopted to tasks in other domains, such as texture description and object proposal generation, but these tasks require annotations for images in the new domain. In this paper, we focus on a novel and challenging task in the pure unsupervised setting: fine-grained image retrieval. Even with image labels, fine-grained images are difficult to classify, letting alone the unsupervised retrieval task. We propose the selective convolutional descriptor aggregation (SCDA) method. The SCDA first localizes the main object in fine-grained images, a step that discards the noisy background and keeps useful deep descriptors. The selected descriptors are then aggregated and the dimensionality is reduced into a short feature vector using the best practices we found. The SCDA is unsupervised, using no image label or bounding box annotation. Experiments on six fine-grained data sets confirm the effectiveness of the SCDA for fine-grained image retrieval. Besides, visualization of the SCDA features shows that they correspond to visual attributes (even subtle ones), which might explain SCDA’s high-mean average precision in fine-grained retrieval. Moreover, on general image retrieval data sets, the SCDA achieves comparable retrieval results with the state-of-the-art general image retrieval approaches.

Anti-Phospholipase A2 Receptor Antibody in Membranous Nephropathy

Abstract: The M-type phospholipase A2 receptor (PLA2R) is a target autoantigen in adult idiopathic membranous nephropathy (MN), but the prevalence of autoantibodies against PLA2R is unknown among Chinese patients with MN. Here, we measured anti-PLA2R antibody in the serum of 60 patients with idiopathic MN, 20 with lupus-associated MN, 16 with hepatitis B (HBV)-associated MN, and 10 with tumor-associated MN. Among patients with idiopathic MN, 49 (82%) had detectable anti-PLA2R autoantibodies using a Western blot assay; an assay with greater sensitivity detected very low titers of anti-PLA2R in 10 of the remaining 11 patients. Using the standard assay, we detected anti-PLA2R antibody in only 1 patient with lupus, 1 with HBV, and 3 with cancer, producing an overall specificity of 89% in this cohort limited to patients with secondary MN. The enhanced assay detected low titers of anti-PLA2R in only 2 additional samples of HBV-associated MN. In summary, these results suggest that PLA2R is a major target antigen in Chinese idiopathic MN and that detection of anti-PLA2R is a sensitive test for idiopathic MN.

Achieving hierarchical hollow carbon@Fe@Fe3O4 nanospheres with superior microwave absorption properties and lightweight features

Abstract: Hierarchical hollow carbon@Fe@Fe3O4 nanospheres were synthesized by a simple template method and another pyrolysis process. Interestingly, the thickness of hollow carbon spheres is tunable by a simple hydrothermal approach. The as-prepared carbon@Fe@Fe3O4 shows excellent microwave absorption properties. In detail, the maximum effective frequency is up to 5.2 GHz with an optimal reflection loss value of −40 dB while the coating thickness is just 1.5 mm. Meanwhile, such absorption properties can be maintained via controlling the thickness of the hollow carbon. For instance, in another coating layer of 2 mm, the effective frequency is still more than 5 GHz as the carbon thickness declines to 12 nm. As novel electromagnetic absorbers, the composites also present the lower density feature due to the hollow carbon sphere frame. The excellent electromagnetic absorption mechanism may be attributed to the obvious interface polarization, and strong magnetic loss ability resulting from the Fe and Fe3O4 shell. Besides, owing to the dielectric feature of carbon, the hollow carbon core is beneficial for the attenuation ability.

Adaptive Neural Output Feedback Control of Uncertain Nonlinear Systems With Unknown Hysteresis Using Disturbance Observer

Abstract: In this paper, an adaptive neural output feedback control scheme is proposed for uncertain nonlinear systems that are subject to unknown hysteresis, external disturbances, and unmeasured states. To deal with the unknown nonlinear function term in the uncertain nonlinear system, the approximation capability of the radial basis function neural network (RBFNN) is employed. Using the approximation output of the RBFNN, the state observer and the nonlinear disturbance observer (NDO) are developed to estimate unmeasured states and unknown compounded disturbances, respectively. Based on the RBFNN, the developed NDO, and the state observer, the adaptive neural output feedback control is proposed for uncertain nonlinear systems using the backstepping technique. The first-order sliding-mode differentiator is employed to avoid the tedious analytic computation and the problem of “explosion of complexity” in the conventional backstepping method. The stability of the whole closed-loop system is rigorously proved via the Lyapunov analysis method, and the satisfactory tracking performance is guaranteed under the integrated effect of unknown hysteresis, unmeasured states, and unknown external disturbances. Simulation results of an example are presented to illustrate the effectiveness of the proposed adaptive neural output feedback control scheme for uncertain nonlinear systems.