Showing papers by "Dorin Comaniciu published in 1996"

Training set synthesis for entropy-constrained transform vector quantization

Abstract: This paper introduces the concept of training set synthesis for entropy-constrained transform vector quantization (TSS-ECTVQ). The statistics of actual sets of transform vectors are first approximated using histograms. New sets of vectors-called synthesized training sets-are obtained based upon the estimated parameters-called training set parameters. By employing a fast entropy-constrained algorithm, codebooks are populated from the synthesized training sets for each image being coded. Then, entropy-constrained vector quantization is performed. The training set parameters are sent to the decoder, which obtains the same training sets, and generates codebooks identical to the encoder. Experimental results demonstrate that high quality image coding at low bit rates can be obtained with the proposed TSS-ECTVQ method. In particular, the image Lenna was coded at 0.25 bits/pixel with a PSNR of 32.42 dB.

Vector quantization clustering using lattice growing search

Abstract: In this paper we introduce a non-iterative algorithm for vector quantization clustering based on the efficient search for the two clusters whose merging gives the minimum distortion increase. The search is performed within the A'-dimensional cells of a lattice having a generating matrix that changes from one step of the algorithm to another. The generating matrix is modified gradually so that the lattice cells grow in volume, allowing the search of the two closest clusters in an enlarged neighborhood. We call this algorithm Lattice Growing Search (LGS) clustering. Preliminary results on 512 × 512 images encoded at 0.5 bits/pixel showed that the LGS technique can produce codebooks of similar quality in less than 1/10 of the time required by the LBG algorithm [9].