Data Research Team Made Progress on Knowledge Graph and Large-scale Remote Sensing Image Data Understanding
The big data research team from the Computer Network Information Center of the Chinese Academy of Sciences made progress on knowledge graph and large-scale remote sensing image data understanding, and relevant results have been accepted by international journals and conferences recently.
Learnable Gated Convolutional Neural Network for Semantic Segmentation in Remote-Sensing Images, submitted by Shichen Guo and Xuezhi Wang, was accepted by Remote Sensing (JCR: Q1; IF: 4.1), which proposed a learning gated convolutional neural network to solve the problem of semantic segmentation of remote sensing images and achieved 93.65% F1 score and 88.06% IoU score in the specific categories of ISPRS datasets.
A Color Enhancement Scene Estimation Approach for Single Image Haze Removal, submitted by Fayaz Ali Dharejo, Yi Du and Yuanchun Zhou, was accepted by IEEE Geoscience and Remote Sensing Letters (JCR: Q1; IF:3.5), which proposed a new and easy way to implement image dehazing method. The experimental results show that in the terms of subjective and visual quality, the color, contrast, naturalness and high brightness of the object increase the image to be improved.
Unsupervised Author Disambiguation using Heterogeneous Graph Convolutional Network Embedding, submitted by Ziyue Qiao, Yi Du and Yuanchun Zhou, was accepted by IEEE International Conference on Big Data. Aiming at the general task of author name disambiguation in academic knowledge graph construction, an unsupervised disambiguation framework based on heterogeneous graph neural network embedding is proposed. Experimental results on two datasets show that the framework is superior to the latest author disambiguation method.