Beihang University

Beijing University of Posts and Telecommunications

University of Science and Technology Beijing

Yifan Zhao

Jia Qi Li

Qinping Zhao

Yue Hua Zhang

Shuang Hu

Xiaowu Chen

Zhiyong Feng

Zhiqing Wei

Qixun Zhang

Sai Huang

Xiao Hu

Yafei Song

Zhaozhao Xiong

Dingfeng Shi

Yaxin Pan

Kaikai Ma

Fulong Dai

Zhuo Kang

Embodiments of the present invention provide a signal processing method and apparatus, a first device determines a target insertion position of a pilot sequence in a first frequency-domain signal based on the data bits in a first sub-data signal obtained by dividing data bits in a data signal, combines the pilot sequences with a second sub-data signal obtained by dividing the data signal according to the determined target insertion positions, to obtain the first frequency-domain signal; sends the pulse signal corresponding to the first frequency-domain signal to the second device. The second device determines the pilot sequences in the first frequency-domain signal corresponding to the pulse signal and determines the data bits corresponding to the insertion position of each of the pilot sequences in the first frequency-domain signal, to obtain the first sub-data signal; demodulates subcarriers in the first frequency-domain signal other than the pilot sequences, to obtain the second sub-data signal; splices the first sub-data signal and the second sub-data signal, to obtain the corresponding data signal. Based on the above processing, spectrum resource utilization rate can be improved.

Signal processing method and apparatus

Embodiments of the present disclosure provides a fine-grained image recognition method and apparatus using graph structure represented high-order relation discovery, wherein the method includes: inputting an image to be classified into a convolutional neural network feature extractor with multiple stages, extracting two layers of network feature graphs in the last stage, constructing a hybrid high-order attention module according to the network feature graphs, and forming a high-order feature vector pool according to the hybrid high-order attention module, using each vector in the vector pool as a node, and utilizing semantic similarity among high-order features to form representative vector nodes in groups, and performing global pooling on the representative vector nodes to obtain classification vectors, and obtaining a fine-grained classification result through a fully connected layer and a classifier based on the classification vectors.

Fine-grained image recognition method and apparatus using graph structure represented high-order relation discovery

A composite aerogel based on graphdiyne motif arrangement and a preparation method thereof are provided by the embodiments of the disclosure, belonging to the technical field of photothermal evaporation materials, including firstly growing graphdiyne on a surface of the graphene oxide to obtain a graphdiyne-coated graphene oxide, taking an additional graphene oxide, adding the additional graphene oxide and the graphdiyne-coated graphene oxide obtained by the coupling reaction into pure water, and ultrasonically mixing to obtain a mixed dispersion, adding a polyvinyl alcohol aqueous solution into the mixed dispersion, and transferring to a reaction kettle after ultrasonic mixing, putting the reaction kettle after sealing into a blast drying box for hydrothermal reaction to obtain a gel structure, and cleaning the gel structure, and freeze-drying to obtain a composite aerogel containing graphdiyne-coated reduced graphene oxide and reduced graphene oxide.

Composite aerogel based on graphdiyne motif arrangement and preparation method thereof

Embodiments of the present invention provide a method and a device for three-dimensional feature-embedded image object component-level semantic segmentation, the method includes: acquiring three-dimensional feature information of a target two-dimensional image; performing a component-level semantic segmentation on the target two-dimensional image according to the three-dimensional feature information of the target two-dimensional image and two-dimensional feature information of the target two-dimensional image. In the technical solution of the present application, not only the two-dimensional feature information of the image but also the three-dimensional feature information of the image are taken into consideration when performing the component-level semantic segmentation on the image, thereby improving the accuracy of the image component-level semantic segmentation.

Method and device for three-dimensional feature-embedded image object component-level semantic segmentation

Growing community of inventors

Beijing, China

Yifan Zhao