ID 原文 译文
56828 本文主要介绍了光电神经突触器件的基本性能和类别,展示了已经实现的光电神经突触器件的模拟应用,并展望了今后光电神经突触器件的发展. In this review we introduce the basic propertiesof optoelectronic synaptic devices. Different types and applications that have been reported for optoelectronicdevices are discussed. In addition, future prospects for the development of optoelectronic synaptic devices isoutlined.
56829 片段视频语义识别旨在识别视频中短小片段的语义概念,是视频分析的一项重要任务. Segment-level video semantic recognition, which known to be an important task in video analysis,attempts to identify the semantic concepts in short video clips.
56830 由于片段视频的数量巨大且缺乏可参考的网络标签,片段视频的标记十分困难,通常只能对部分片段视频进行标记. Labeling video segments is difficult because thereis an extremely large number of segments and there are no network tags; consequently, only a portion of the videosegments are labeled.
56831 如何利用有限的语义标签提高片段视频语义识别的准确率是一项关键挑战. Determining how to improve the accuracy of semantic recognition of fragmented videos withlimited semantic labels is a key challenge in video semantic recognition.
56832 因此本文提出了一种基于长短时预测一致性的视频语义识别算法. This paper proposes a video semanticrecognition algorithm based on the consistency of video- and segment-level predictions.
56833 该算法通过引入完整视频语义与片段视频语义一致性的约束,对片段视频语义识别结果进行筛选,以此提高片段视频语义识别的准确率. The proposed algorithmintroduces the constraint of consistency between complete video semantics and fragmentary video semantics. Theproposed algorithm can be applied to filter the video segment semantic results to improve recognition accuracy.
56834 本文提出的算法在大规模视频数据集YouTube-8M的片段视频语义识别任务上达到了82.62%的平均均值准确率(mean average precision, MAP)识别精度,在第三届YouTube-8M比赛中排名第二. The proposed algorithm achieved 82. 62% mean average precision on the video segment semantic recognition taskusing the large-scale video dataset YouTube-8M and ranked second in the third YouTube-8M competition
56835 近年来,视频数据资源的日益丰富催生了一系列对于视频片段精细检索的需求. In recent years, increasing amounts of video resources have created a series of demands for fine retrievalof video moments, such as highlight moments in sports events and the re-creation of specific video content.
56836 在这样的背景下,对于跨模态视频片段检索的研究逐渐兴起,其旨在根据输入的查询文本,输出一段视频中符合文本描述的片段. Inthis context, research on cross-modal video segment retrieval, which attempts to output a video moment thatmatches the input query text, is gradually emerging.
56837 现有的研究工作主要关注于查询文本与视频片段的全局或局部的特征表达,而忽略了查询文本与视频片段中所蕴含的语义关系在跨模态检索中的匹配. Existing solutions primarily focus on global or local featurerepresentation for query text and video moments. However, such solutions ignore matching semantic relationscontained in query text and video moments.