-
公开(公告)号：KR101069602B1
公开(公告)日：2011-10-05
申请号：KR1020090092515
申请日：2009-09-29
Applicant: 한국과학기술원
Abstract: 또한, 물체인식시스템에있어서는전체입력영상의글로벌특징인물체의관심영역을뉴로-퍼지시스템이빨리찾아내고, 이를기반으로고성능병렬처리프로세서로선택된영역에한해서물체의로컬특징을추출하는자세한이미지를처리함으로써, 영상처리속도의가속화, 저면적, 및저 전력으로실시간물체인식을가능하게한다.
-
公开(公告)号：KR1020110034991A
公开(公告)日：2011-04-06
申请号：KR1020090092515
申请日：2009-09-29
Applicant: 한국과학기술원
Abstract: PURPOSE: A computer system connecting a neuro-fuzzy system with parallel processor and an object recognizing system using the same are provided to achieve the calculation quantity reduction and consumed power reduction of a neural network and purge network. CONSTITUTION: A neuro-fuzzy system(1100) includes an analog circuit unit, a digital circuit unit, and a conversion circuit unit. The conversion circuit unit converts each data value of the analog circuit unit and the digital circuit unit into domain information. The neuro-fuzzy system performs to processing to create global information for input data, and a parallel processor(1200) performs the processing to create local information for the global information.
Abstract translation: 目的：提供将神经模糊系统与并行处理器连接的计算机系统和使用其的物体识别系统，以实现神经网络和净化网络的计算量减少和功耗降低。 构成：神经模糊系统（1100）包括模拟电路单元，数字电路单元和转换电路单元。 转换电路单元将模拟电路单元和数字电路单元的每个数据值转换为域信息。 神经模糊系统进行处理以创建用于输入数据的全局信息，并行处理器（1200）执行用于为全局信息创建本地信息的处理。
-
Search Results
2
records
(took 0.017 seconds)
