Hongguang Zhang

Hongguang Zhang

Auto-generated profile, based on publicly available information

Hongguang Zhang

Systems Engineering Institute, AMS, Beijing, China

Research Focus & Metrics

Academic Career Length

9Years (2018-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
29 October 2020
Verified
Book Chapter
Closed access
PDF file
HZHongguang Zhang
Hongguang Zhang
LZLi Zhang
Li Zhang
XQXiaojuan Qi
Xiaojuan Qi
HLHongdong Li
Hongdong Li
PTPhilip H. S. Torr
Philip H. S. Torr
PKPiotr Koniusz
Piotr Koniusz

24 August 2021 in IEEE Transactions on Pattern Analysis and Machine Intelligence
Verified
Journal Article
Closed access
PKPiotr Koniusz
Piotr Koniusz
HZHongguang Zhang
Hongguang Zhang
Power Normalizations ( PN ) are useful non-linear operators which tackle feature imbalances in classification problems. We study PNs in the deep learning setup via a novel PN layer pooling feature maps. Our layer combines the feature vectors and their respective spatial locations in the feature maps produced by the last convolutional layer of CNN into a positive definite matrix with second-order statistics to which PN operators are applied, forming so-called Second-order Pooling ( SOP ). As the main goal of this paper is to study Power Normalizations, we investigate the role and meaning of MaxExp and Gamma, two popular PN functions. To this end, we provide probabilistic interpretations of such element-wise operators and discover surrogates with well-behaved derivatives for end-to-end training. Furthermore, we look at the spectral applicability of MaxExp and Gamma by studying Spectral Power Normalizations ( SPN ). We show that SPN on the autocorrelation/covariance matrix and the Heat Diffusion Process (HDP) on a graph Laplacian matrix are closely related, thus sharing their properties. Such a finding leads us to the culmination of our work, a fast spectral MaxExp which is a variant of HDP for covariances/autocorrelation matrices. We evaluate our ideas on fine-grained recognition, scene recognition, and material classification, as well as in few-shot learning and graph classification.

13 January 2022 in IEEE Transactions on Multimedia
Verified
Journal Article
Closed access
HZHongguang Zhang
Hongguang Zhang
HLHongdong Li
Hongdong Li
PKPiotr Koniusz
Piotr Koniusz
We propose a Multi-level Second-order (MlSo) few-shot learning network for supervised or unsupervised few-shot image classification and few-shot action recognition. We leverage so-called power-normalized second-order base learner streams combined with features that express multiple levels of visual abstraction, and we use self-supervised discriminating mechanisms. As Second-order Pooling (SoP) is popular in image recognition, we employ its basic element-wise variant in our pipeline. The goal of multi-level feature design is to extract feature representations at different layer-wise levels of CNN, realizing several levels of visual abstraction to achieve robust few-shot learning. As SoP can handle convolutional feature maps of varying spatial sizes, we also introduce image inputs at multiple spatial scales into MlSo. To exploit the discriminative information from multi-level and multi-scale features, we develop a Feature Matching (FM) module that reweights their respective branches. We also introduce a self-supervised step, which is a discriminator of the spatial level and the scale of abstraction. Our pipeline is trained in an end-to-end manner. With a simple architecture, we demonstrate respectable results on standard datasets such as Omniglot, mini-ImageNet, tiered-ImageNet, Open MIC, fine-grained datasets such as CUB Birds, Stanford Dogs and Cars, and action recognition datasets such as HMDB51, UCF101, and mini-MIT.

11 August 2020 in IEEE Access
Verified
Journal Article
Open access
ZZZhang Zhang
Zhang Zhang
HZHongguang Zhang
Hongguang Zhang
JCJihua Chen
Jihua Chen
LWLei Wang
Lei Wang
Despite CNN-based deblur models have shown their superiority when solving motion blurs, restoring a photorealistic image from severe motion blur remains an ill-posed problem due to the loss of temporal information and textures. Event cameras such as Dynamic and Active-pixel Vision Sensor (DAVIS) can simultaneously produce gray-scale Active Pixel Sensor (APS) frames and events, which can capture fast motions as events of very high temporal resolution, i. e. , $1~\mu s$ , can provide extra information for blurry APS frames. Due to the natural noise and sparsity of events, we employ a recurrent encoder-decoder architecture to generate dense recurrent event representations, which encode the overall historical information. We concatenate the original blurry image with the event representation as our hybrid input, from which the network learns to restore the sharp output. We conduct extensive experiments on GoPro dataset and a real event blurry dataset captured by DAVIS240C. Our experimental results on both synthetic and real images demonstrate state-of-the-art performance for $1280\times 720 $ images at 30 fps.

14 November 2022 in International Journal of Computer Vision
Verified
Journal Article
Closed access
PDF file
HZHongguang Zhang
Hongguang Zhang
LZLimeng Zhang
Limeng Zhang
YDYuchao Dai
Yuchao Dai
HLHongdong Li
Hongdong Li
PKPiotr Koniusz
Piotr Koniusz

10 July 2024
Verified
Conference Paper
Open access
JWJiandong Wang
Jiandong Wang
HZHongguang Zhang
Hongguang Zhang
Chun LiuChun Liu
Chun Liu
XYXiongjun Yang
Xiongjun Yang
The effective utilization of multimodal information is crucial for the task of fake news detection. The primary challenge lies in reducing the semantic distance between the news evidences of different modalities and accurately aligning them in the latent space. It has been observed that the relevance of news image content varies across coarse-to-fine scales, depending on the object and its class label. For instance, the visual evidence to justify a news item sometimes may be a very local detail such as the tiny changes of facial expression and limb position while sometimes may be the global composition. Consequently, how to align news text with image at the most discriminative scale significantly impacts detection performance. However, very few studies have addressed this issue in fake news detection. In this paper, we delve deeper into this issue and propose a simple yet effective Multi-scale Semantic Alignment and Cross-modal Attention (MSACA) network. Specifically, we construct hierarchical multi-scale images for each news data, enhance the semantic consistency between text and images in the latent space, and employ an attention module to select the deterministic embeddings in an end-to-end manner. Extensive experiments on two real-world benchmarks demonstrate the superior performance of our proposed MSACA network.