Name
Abstract | ||
|---|---|---|
Data compression is an important component of communication and storage technologies, directly contributing to the efficiency of these systems. Lossy data or signal compression, while not applicable to all applications, usually allows for much higher efficiency. In this paper we propose a method for adopting the concept of mixed- transforms and evaluate it for vehicle trajectory data. In mixed- transform methods, compression residuals from the first representation of a signal are passed to a second transformation stage where further compression is applied in that domain. The final residual which represents error resulting from the compression and decompression process, is then used as a cost to be minimized by adjusting the configuration of each transformation stage. An adaptive algorithm from earlier works is adopted to iteratively adjust the configuration of each compression stage (selection of coefficients from each transform) to minimize the error. A hierarchical realization of the iterative adaptive algorithm is presented that uses preset compression ratios and optimizes the coefficient selection accordingly. We show that using DCT followed by Haar in the proposed method, it is possible to efficiently compress vehicle trajectory data. The results show that a higher quality for reconstructed data can be achieved using this method, compared to using a single transform compression. For trajectories with sudden movements, the method has a more notable improvement. |
Year | DOI | Venue |
|---|---|---|
2018 | 10.1109/mwscas.2018.8623987 | Midwest Symposium on Circuits and Systems Conference Proceedings |
Keywords | Field | DocType |
signal compression,DCT,Haar,transforms,mixed transforms,nonorthogonal basis | Residual,Lossy compression,Computer science,Discrete cosine transform,Algorithm,Electronic engineering,Compression ratio,Adaptive algorithm,Data compression,Trajectory,Signal compression | Conference |
ISSN | Citations | PageRank |
1548-3746 | 0 | 0.34 |
References | Authors | |
0 | 2 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Masoumeh Kalantari Khandani | 1 | 1 | 1.40 |
| Wasfy B. Mikhael | 2 | 76 | 76.27 |