发明名称 IMAGE ENCODING METHOD
摘要 PROBLEM TO BE SOLVED: To reduce the number of calculation as addition/subtraction by discriminating whether the respective components of two vectors as the objects of simplified norm calculation are mutually sufficiently close or not while using a simplified norm calculation expression eliminating multiplication from norm calculation. SOLUTION: An image vector Xi is inputted and the initial setting of k=0, t=0 and Nmin=∞ is performed (s1 and s2). The value of (k) is increased just by 1 and a reference value α=αk is calculated (s3 and s4). The value of (t) is increased just by 1, a representative vector Yt is inputted and approximate discrimination is performed (s5-s7). A norm N* is calculated by the required expression (s8). It is discriminated whether Nmin>Nt* is established or not (s9). Since Nmin>Nt is established first, the conditions of Nmin=Nt* and -capped vector Xi = representative vector Yt are established (s10). When it is discriminated the value of (t) gets equal with a maximum value (s) of a code (j), it is discriminated whether Nmin=∞ is established or not and the code (j) showing the final -capped vector Xi = representative vector Yj is outputted (s11, s12 and s13).
申请公布号 JPH1188875(A) 申请公布日期 1999.03.30
申请号 JP19970236774 申请日期 1997.09.02
申请人 MATSUSHITA ELECTRIC IND CO LTD 发明人 SHIMODA HARUAKI;KOBAYASHI MASAAKI;KONO AKIHIRO;NOMA MUTSUAKI
分类号 H04N19/126;H03M7/30;H04N1/41;H04N7/24;H04N19/00;H04N19/136;H04N19/176;H04N19/186;H04N19/196;H04N19/423;H04N19/94 主分类号 H04N19/126
代理机构 代理人
主权项
地址