디지털 포렌직 기술을 이용한 보안 인쇄물 위조식별 시스템 및 그 방법
    1.
    发明授权
    디지털 포렌직 기술을 이용한 보안 인쇄물 위조식별 시스템 및 그 방법 有权
    基于数字法医技术的安全打印资源伪造检测系统及其方法

    公开(公告)号:KR101140699B1

    公开(公告)日:2012-05-03

    申请号:KR1020100052297

    申请日:2010-06-03

    Abstract: PURPOSE: The forgery detection system of a security printed material and a detection method thereof are provided to create learning model for detecting forgery by performing machine learning with extracted feature points. CONSTITUTION: The forgery detection system of a security printed material comprises an image transformation part(10), a feature point extracting part(20), a machine learning part(30), and a forgery detection part(40). The image transformation part changes prints into a digital. The feature point extracting part extracts the feature point of the digital image. The machine learning part performs machine learning using the feature points. The forgery detection part detects the forgery of a printed material from a machine learning result.

    디지털 포렌직 기술을 이용한 보안 인쇄물 위조식별 시스템 및 그 방법
    2.
    发明公开
    디지털 포렌직 기술을 이용한 보안 인쇄물 위조식별 시스템 및 그 방법 有权
    基于数字审查技术的安全打印资源的格栅检测系统及其方法

    公开(公告)号:KR1020110132770A

    公开(公告)日:2011-12-09

    申请号:KR1020100052297

    申请日:2010-06-03

    CPC classification number: G07D7/2016 G06T1/00

    Abstract: PURPOSE: The forgery detection system of a security printed material and a detection method thereof are provided to create learning model for detecting forgery by performing machine learning with extracted feature points. CONSTITUTION: The forgery detection system of a security printed material comprises an image transformation part(10), a feature point extracting part(20), a machine learning part(30), and a forgery detection part(40). The image transformation part changes prints into a digital. The feature point extracting part extracts the feature point of the digital image. The machine learning part performs machine learning using the feature points. The forgery detection part detects the forgery of a printed material from a machine learning result.

    Abstract translation: 目的:提供安全印刷材料的伪造检测系统及其检测方法,通过提取特征点进行机器学习,创建用于检测伪造的学习模型。 构成:安全印刷材料的伪造检测系统包括图像变换部分(10),特征点提取部分(20),机器学习部分(30)和伪造检测部分(40)。 图像变换部分将打印件更改为数字。 特征点提取部分提取数字图像的特征点。 机器学习部分使用特征点进行机器学习。 伪造检测部件通过机器学习结果检测印刷品的伪造。

    샷 기반의 센서 패턴 노이즈를 이용한 불법 포르노그래피 탐지 방법 및 그 장치
    3.
    发明授权
    샷 기반의 센서 패턴 노이즈를 이용한 불법 포르노그래피 탐지 방법 및 그 장치 有权
    使用基于拍摄的传感器图案噪声的非法图像视频检测

    公开(公告)号:KR101518132B1

    公开(公告)日:2015-05-07

    申请号:KR1020130134657

    申请日:2013-11-07

    CPC classification number: H04N5/357 G06T7/0002 G06T2207/10016

    Abstract: 샷기반의센서패턴노이즈를이용한불법포르노그래피탐지방법및 그장치가개시된다. 영상판별방법은, 입력영상을적어도하나의프레임으로이루어진프레임집합으로분류하는단계; 상기프레임집합각각에대하여상기프레임집합으로부터센서패턴노이즈(sensor pattern noise)를추정하는단계; 상기센서패턴노이즈로부터상관도를계산하는단계; 및상기상관도에따라상기입력영상에대한불법영상여부를판별하는단계를포함할수 있다.

    Abstract translation: 公开了一种通过使用基于射击的传感器图案噪声来检测非法色情内容的方法及其装置。 在本发明中,用于识别图像的方法包括以下步骤:将输入图像分类成由至少一个帧组成的帧组; 从每个帧组的帧集合估计传感器模式噪声; 从传感器模式噪声计算相关性; 以及基于所述相关性来确定所述输入图像是否是非法的。

Patent Agency Ranking