| 摘要 |
<p>Data compression using vector quantization is performed at a high speed by using hardware. Among a plurality of template patterns, a template pattern most similar to the input pattern is searched. For this, a template pattern calculating similarity by using feature amounts of the input pattern and the template pattern is selected. When calculating the similarity between the selected template pattern and the input pattern, the calculation is performed by a bit serial way so as to reduce the number of templates requiring matching by a pattern matching processing in the vector quantization and reduce the number of cycles required for calculation of the matching, thereby executing the data compression using the vector quantization at a high speed.</p> |
