ID 原文 译文
963 实验结果表明，本文方法具有较高的精度和鲁棒性，尤其对于复杂场景、运动遮挡和运动边缘模糊的图像具有较好的边缘保护作用。 The experimental results demonstrate that our method has high accuracy and good ro-bustness, and especially has significant benefit of boundary preserving in the areas of complex scene, motion occlusion andmotion boundary.
964 图模型越来越广泛地应用于数据管理、知识发现和信息服务等问题中， Graphs are increasingly used in data management, knowledge discovery and information services.
965 图嵌入作为图分析和应用的重要技术手段，成为了人工智能领域研究的热点之一。 As animportant strategy of graph analysis and applications, graph embedding has become one of the subjects with great attention inartificial intelligence.
966 本文从图嵌入研究中面临的挑战出发，主要介绍了基于矩阵分解、基于随机游走和基于深度学习的图嵌入方法。 Starting from the challenges faced in graph embedding studies, this paper introduces the principal meth-ods based on matrix decomposition, random walk and deep learning.
967 接着，介绍了图嵌入方法常用的测试数据集、评测标准和典型应用。 Then, we introduce general test datasets, evaluation cri-teria as well as typical applications widely used in graph embedding.
968 最后，总结了图嵌入未来研究的趋势和方向。 Finally, we summarize the trend and future research is-sues of graph embedding.
969 视频目标分割是计算机视觉领域中的一个研究热点，传统基于深度学习的视频目标分割方法在线微调深度网络，导致分割耗时长，难以满足实时的需求。 Video object segmentation (VOS)is a research hotspot in the field of computer vision. Traditional VOS based on deep learning fine-tunes the deep network online, which leads to long time-consuming segmentation and is difficultto meet real-time requirements.
970 本文提出一种快速的视频目标分割方法。 Therefore, we propose a fast VOS method.
971 首先，参数共享的孪生编码器子网将参考流和目标流映射到相同的特征空间，使得相同的目标具有相似的特征。 First, the weight-shared siamese encoder subnet maps the reference stream and the target stream to the same feature space;so that the same objects have similar features.
972 然后，全局特征提取子网在特征空间中匹配给定目标相似的特征，定位目标对象。 Then, the global feature extraction subnet matches the features similar to the given object to locate the object.