Title :
Tree encoding of line drawings
Author :
Vembar, Manindranath ; Mohan, Seshadri
Author_Institution :
Identification & Security Syst. Inc., Albany, NY, USA
fDate :
9/1/1988 12:00:00 AM
Abstract :
Recently, tree-searched chain codes have been proposed that reduce the reconstruction distortion significantly compared to chain codes. The authors study tree coding schemes to encode line drawings and show that they outperform the basic and tree-searched chain codes. Both adaptive and nonadaptive tree coding schemes are studied. The adaptive schemes considered include an adaptive gradient algorithm and a Kalman filtering algorithm, both in conjunction with tree coding. The adaptive tree encoding schemes outperform all other schemes considered
Keywords :
encoding; filtering and prediction theory; trees (mathematics); Kalman filtering; adaptive gradient algorithm; chain codes; line drawings; tree coding; Acoustic distortion; Acoustic signal processing; Convolution; Convolutional codes; Encoding; Engineering drawings; Filtering algorithms; Multidimensional signal processing; Quantization; Speech processing;
Journal_Title :
Acoustics, Speech and Signal Processing, IEEE Transactions on