• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>2015 Global Conference on Communication Technologies >Face identification methodologies in videos
【24h】

Face identification methodologies in videos

机译：视频中的人脸识别方法

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Face recognition is an evolving area, changing and improving constantly. Face recognition through video has recently drawn lots of attention from both the research community and industry and is beginning to be applied in a variety of domains, predominantly for security. Searching image in dictionary which consists of large number of video frames is a challenging task. Various techniques have been introduced in the past few years which use the technique to process video frames in serial manner which results to lack of performance degradation. We proposed method with which recognition can now be carried out to process video frames in a parallel manner and with this parallel processing of video frame's rank list a substantial amount of time would help to increase the efficiency and decrease the execution time. An approach is proposed to increase the efficiency of the system in which video frames are divided into four parts to increase the speed of the system. The images obtained are then ranked, clustered and re-ranked to get the matching images of the two videos. The efficacy of proposed system can be evaluated using some standard database such as You Tube or MBGC v2 database and shows better results as compared to existing methods. The proposed approach works for each video frames in a sequence. For a sequence of frames, the likelihoods are summed, and compared at the end of the sequence, taking the maximum likelihood training model as the correct result. With this approach we illustrate that our technique is efficient and performs extensively better than many viable video-based face recognition algorithms.
机译：人脸识别是一个不断发展的领域，不断变化和改进。通过视频进行面部识别最近引起了研究界和行业的广泛关注，并且已开始在各种领域中得到广泛应用，主要是出于安全性考虑。在包含大量视频帧的字典中搜索图像是一项艰巨的任务。在过去几年中已经引入了各种技术，这些技术使用该技术以串行方式处理视频帧，从而导致性能降低。我们提出了一种方法，现在可以使用该方法以并行方式对视频帧进行识别，并且通过并行处理视频帧的等级列表，大量时间将有助于提高效率并减少执行时间。提出了一种提高系统效率的方法，其中将视频帧分为四个部分以提高系统速度。然后对获得的图像进行排名，聚类和重新排名，以获得两个视频的匹配图像。可以使用某些标准数据库（例如You Tube或MBGC v2数据库）来评估所建议系统的功效，并且与现有方法相比，该系统显示出更好的结果。所提出的方法适用于序列中的每个视频帧。对于帧序列，将似然相加，并在序列末尾进行比较，以最大似然训练模型为正确结果。通过这种方法，我们证明了我们的技术是有效的，并且比许多基于视频的可行人脸识别算法具有更好的性能。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号