首页> 外国专利> METHOD, APPARATUS, AND PROGRAM FOR SAMPLING LEARNING TARGET FRAME IMAGE OF VIDEO FOR AI IMAGE LEARNING, AND METHOD OF SAME IMAGE LEARNING

METHOD, APPARATUS, AND PROGRAM FOR SAMPLING LEARNING TARGET FRAME IMAGE OF VIDEO FOR AI IMAGE LEARNING, AND METHOD OF SAME IMAGE LEARNING

机译:用于对用于AI图像学习的视频的学习目标帧图像进行采样的方法,装置和程序,以及用于相同图像学习的方法

摘要

To provide a method, an apparatus, and a program for sampling a learning target frame image of a video for AI image learning, and a method for the image learning.SOLUTION: A sampling method comprises the steps of: receiving an original video; extracting a predetermined number of frame images per predetermined time interval from the original video; detecting one or more predetermined types of one or more learning target objects on each frame image; removing a background on each frame image; measuring an amount of movement of each of the learning target objects on the n-th (n is a natural number equal to or greater than 2) frame image from which the background is removed; and selecting the n-th frame image as a learning target frame image by comparing a measurement result of the amount of movement of each of the one or more learning target objects detected on the n-th frame image with a predetermined criterion.SELECTED DRAWING: Figure 1
机译:提供一种用于对用于AI图像学习的视频的学习目标帧图像进行采样的方法,装置和程序以及用于图像学习的方法。解决方案:采样方法包括以下步骤:接收原始视频;以及每预定时间间隔从原始视频中提取预定数量的帧图像;在每一帧图像上检测一个或多个预定类型的一个或多个学习目标对象;在每个帧图像上去除背景;测量在去除背景的第n个(n是等于或大于2的自然数)帧图像上的每个学习目标对象的移动量;通过将在第n帧图像上检测到的一个或多个学习目标对象的每个运动量的测量结果与预定标准进行比较,选择第n帧图像作为学习目标帧图像。图1

著录项

  • 公开/公告号JP2020052999A

    专利类型

  • 公开/公告日2020-04-02

    原文格式PDF

  • 申请/专利权人 CROWDWORKS INC;

    申请/专利号JP20190012692

  • 发明设计人 PARK MINWOO;

    申请日2019-01-29

  • 分类号G06T7;G06N20;

  • 国家 JP

  • 入库时间 2022-08-21 11:36:39

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