Loading...
Please wait, while we are loading the content...
Similar Documents
Linear (zero-one) programming approach to fixed-rate entropy-coded vector quantisation
| Content Provider | Semantic Scholar |
|---|---|
| Author | Khandani, Amir K. |
| Copyright Year | 1999 |
| Abstract | The problem of the decoding of a shaped set is formulated in terms of a zero-one linear program. Some special features of the problem are exploited to relax the zero-one constraint, and to substantially reduce the complexity of the underlying simplex search. The proposed decoding method has applications in fixed-rate entropy-coded vector quantisation of a memoryless source, in decoding of a shaped constellation, and in the bit allocation problem. The first application is considered and numerical results are presented for the quantisation of a memoryless Gaussian source demonstrating substantial (of the order of a few tens to a few hundred times) reduction in the complexity with respect to the conventional methods based on dynamic programming. It is generally observed that the complexity of the proposed method has a linear increase with respect to the quantiser dimension. The corresponding numerical results show that it is possible to get very close to the bounds determined by the rate-distortion theory, while keeping the complexity at a relatively low level. |
| Starting Page | 275 |
| Ending Page | 282 |
| Page Count | 8 |
| File Format | PDF HTM / HTML |
| DOI | 10.1049/ip-com:19990133 |
| Volume Number | 146 |
| Alternate Webpage(s) | http://www.cst.uwaterloo.ca/j/AKCOMM99.pdf |
| Alternate Webpage(s) | https://doi.org/10.1049/ip-com%3A19990133 |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
