首页> 外国专利> Compressing apparatus and method of trained deep artificial neural networks for video coding

Compressing apparatus and method of trained deep artificial neural networks for video coding

机译:培训深层人工神经网络进行视频编码的压缩装置和方法

摘要

Disclosed is an apparatus for compressing a trained deep artificial neural network for video coding. The compression apparatus according to an embodiment includes a pruning unit for selectively pruning a trained deep neural network of a coding unit constituting video coding, and a quantization for quantizing the trained deep neural network pruned by the pruning unit. And an entropy coding unit for entropy coding the trained deep neural network quantized by the quantization unit and outputting a bitstream. At this time, the pruning unit calculates the value of the elements describing the trained deep neural network, and calculates the value by reflecting the propagation backward through the deep neural network.
机译:公开了一种用于压缩训练的深层人工神经网络以进行视频编码的装置。根据实施例的压缩装置包括修剪单元,用于选择性地修剪构成视频编码的编码单元的训练的深神经网络,以及用于量化由修剪单元修剪的训练的深神经网络的量化。和熵编码单元,用于熵编码由量化单元量化的训练的深神经网络并输出比特流。此时,修剪单元计算描述训练的深神经网络的元素的值,并通过深神经网络反映向后传播来计算值。

著录项

  • 公开/公告号KR20210035678A

    专利类型

  • 公开/公告日2021-04-01

    原文格式PDF

  • 申请/专利权人 (주)인시그널;

    申请/专利号KR1020190117793

  • 发明设计人 천승문;고현철;

    申请日2019-09-24

  • 分类号G06N3/08;G06N3/04;H04N19/124;H04N19/13;

  • 国家 KR

  • 入库时间 2022-08-24 18:06:46

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

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