A new fast codeword search algorithm for imagc vcctor quantisation (VQ) is introduced. This algorithm performs a fast codeword search in the Hadamard transform (HT) domain using the partial distance search (PDS) technique, Experimental results show that the algorithm needs only 2-3% of the distortion calculations of the exhaustive search method.

Summary A novel fast KNN classification algorithm is proposed for pattern recognition. The technique uses one important feature, mean of the vector, to reduce the search space in the wavelet domain. Since the proposed algorithm rejects those vectors that are impossible to be the k closest vectors in the design set, it largely reduces the classification time and holds the classification performance as that of the original classification algorithm. The simulation on texture image classification confirms the efficiency of the proposed algorithm.

In this paper, a new and fast encoding algorthm for vector quantization is presenfed. This algorithm makes fall use of two characterisstics of a vector: the sum and the variance. A vector is separated into two subvectors: one is composed of the first half of vector components and the other consists of the remaining vector components. Three inequalities based on the sums and varances of a vector and Its two subyectors components are introduced to rejict those codewords that are impossible to be the nearest code-word, thereby saving a great deal of computational time, while in-troducing no estra distortion compared to the conventional full search algorithm. The simulation results show that the proposed algorithm is faster than the equal-average nearest neighbor search (ENNS), the improved ENNS, the equal-average equal-varlance nearest neighbor search (EENNS)and the improved EENNS algo-rithms. Comparing with the improved EENNS algorithm, the pro-posed algorithm reduces the computational time and the number of distortion calculations by 2.4% to 6%and 20.5% to 26.8%. re-spectively The average improvements of the computational time and the number of distortion calculations are 4% and 24.6% for the codebook sizes of 128 to 1024, respectively.

Block truncation coding (BTC) is an efficient block coding technique suitable for real -time image compression, and it has high channel error re-sisting capability and good reconstructed image qual-ity. Abaolute moment BTC (AMBTC) is a simple and fast variant of BTC. The main shortcoming of the original AMBTC algorithm is the high bit rate (nor-mally 2bits/pixel). In order to reduce the bit rate of AMBTC, an efficient BTC image compression algo-rithm with vector quantization (VQ) is presented in this paper. The main idea of the proposed algorithm is to reduce the number of bits required to code the higher mean, the lower mean and the bit plane, which are created by AMBTC for each image block. On the one hand, a simple look-up-table method is presented for coding the higher mean and the lower mean of a block. On the other hand, vector quantization tech-nique is introduced to reduce the number of bits used to code the bit plane. Test results prove the effective-ness of the proposed algorithm.

Codeword index assignment (CIA) is a key issue to vector quantization (VQ) In the communication system with channel crrors that will introduce extra distortions in the decoding step. Tabu search algorithm (TSA) has been success-fully used to solve the codeword index assignment for the purpose of minimizing the extra distortions due to blt errors. In this paper, simulated annealing (SA) technique is introduced In the iteration of TSA to Improve the convergence performance, Experimental results show that the modified tabu search algorithm (MTSA) is superior to TSA by evaluating the performance of channel distortion after the same number of iterations.