Abstract:
본 발명은 멀티모달 및 멀티스케일 특징 기반 유해 영상 차단 장치 및 방법에 관한 것이다. 본 발명에 따른 멀티모달 및 멀티스케일 특징 기반 유해 영상 차단 장치는, 영상 학습 데이터로부터 추출되는 멀티모달 정보를 분석하여 멀티스케일 유해 및 무해 특징을 생성하는 멀티스케일 특징 분석부; 상기 생성된 멀티스케일 유해 및 무해 특징을 통계 처리하고 기계 학습을 수행하여 다수준 유해 분류 모델을 생성하는 유해 분류 모델 생성부; 유해성 판별을 위해 입력되는 입력 영상 데이터로부터 추출되는 멀티모달 정보를 분석하여 상기 입력 영상의 멀티스케일 특징중 적어도 하나를 추출하고 상기 추출된 특징을 상기 다수준 유해 분류 모델중 적어도 하나와 비교함으로써 상기 영상의 유해성을 판단하는 유해성 판단부; 및 상기 판단 결과가 유해인 경우 상기 입력 영상을 차단하는 유해 영상 차단부를 포함한다.
Abstract:
The present invention relates to a method and an apparatus for providing a concealed software execution environment based on virtualization which supports a secure element. Important security information provided by the secure element can be protected from being exposed by an unauthorized access when used in an actual application by configuring a domain which is operated in a closed manner, not exposed to the outside, using a domain isolation technique based on virtualization and by executing the security information such as key information provided by the secure element only within the concealed domain. [Reference numerals] (400) Secure element unit; (410) Secure element execution unit; (420) Command processing unit; (500) Opened type operating environment unit; (600) Concealed type operating environment unit; (610) Security service request management unit; (620) Security service execution unit; (630) Session management unit; (640) Secure element management unit; (650) Command transmission unit; (700) Hypervisor; (AA) Application program; (BB) Security service requesting unit
Abstract:
PURPOSE: A harmful video classifying method and an apparatus thereof are provided to apply general streaming service content and to supply a video determining framework. CONSTITUTION: A second statistic combining unit(750) combines a first combining value to a video clip unit. The second statistic combining unit generates a second combining value. A third statistic combining unit(760) combines the second combining value. The third statistic combining unit generates the final combining value. A video harmful determining unit(770) determines whether to exist intended characteristic according to the predefined determining standard about a video through the final combining value. [Reference numerals] (210) Video; (270) First combining value; (280) Second combining value; (290) Final combining value; (310) Time color histogram feature; (330) Time color energy feature; (350) Time motion energy feature; (710) Video clip and segment dividing unit; (720) Video feature extracting unit; (730) Video feature harmful/harmless determining unit; (740) First statistic combining unit; (750) Second statistic combining unit; (760) Third statistic combining unit; (770) Video harmful/harmless determining unit
Abstract:
PURPOSE: A system for safe content services and method therefor are provided to protect teenagers from harmful content by enabling the teenagers to safely use content trough a portable terminal. CONSTITUTION: When new content is created, a first portable terminal(100) extracts the meta information of the content. A server(200) receives the meta information of the content from the first terminal. The server outputs analysis results by analyzing the content based on file names and hash values related to the content including the meta information.