| 摘要 |
<p>PURPOSE:To favorably compress picture data by dividing original picture data carrying a two-dimentional picture into blocks consisting of plural standards, and vector-quantity quantizing the mean value and, for instance, the binarized data of every block. CONSTITUTION:The original picture data S, showing one sheet of continuous pictures, is made to pass through a block conversion circuit 9 at first, and is converted into the data Sb of every rectangular block, consisting of prescribed picture elements. The data Sb of every block is encoded by a block encoding circuit 10. In this circuit 10, the mean value S0 of every block B is obtained, and the data, larger than S0, is made to be '1', and the data, smaller than SO, is made to be '0', and the mean values m1, m2 if the '0' value and the '1' value are obtained, and the circuit 10 sends the data M to show the mean values m1, m2 and the binarized data (x), to a vector quantization circuit 11. The circuit 11 selects a vector, corresponding to the group of the binarized data and the minimum distortion of every inputted block, out of the plural vectors, previously stored in a memory 12 as a code book. Then, the picture, nearly equivalent to the original picture, is reproduced from a picture reproducing device 17.</p> |
