Quantization

Adaptive quantization using piecewise companding and scaling for Gaussian mixture

“…The LloydMax quantizer for SGMM is found by the LBG algorithm numerically [2]. The uniform quantizer we compare with is the optimal uniform quantizer which is applied uniformly to the finite region containing 99.8% of the data of the GMM distribution.…”

“…Lloyd-Max quantizers and companders are already well developed for Gaussian distribution or Laplacian distribution [2,4,5] as convenience, but not for Gaussian mixture model (GMM). Since GMM serves as a good approximation of an arbitrary distribution, it is important to develop quantizers and companders for GMM, which are expected to find wide applications in ADC and high dynamic range (HDR) image compression, as well as audio [6] and video [7] compression.…”

“…Existing quantization schemes can be classified into two categories, namely, uniform quantization and nonuniform quantization [2,3]. Uniform quantization is simple, but not optimal for signals with nonuniform distribution in terms of MMSE if more computations and storage are available.…”

Construção de Dicionários Voltados para a Redução da Complexidade Computacional da Etapa de Codificação da Quantização Vetorial

“…Em outras palavras, essa teoria ressalta a superioridade da quantização vetorial (QV) sobre a quantização escalar [4]. Matematicamente, a quantização vetorial pode ser definida como um mapeamento ¥ de um vetor de entrada [7,8]), é expressa em bit/amostra.…”

Chatting in distributed quantization networks

“…Here, performance results from Shannon theory are overly optimistic since tools such as joint-typicality encoding and decoding are not reliable. Instead, we consider the complementary asymptotic of high resolution, where the blocklength is considered fixed and the compression rate R is made large [17], [18]. We now introduce high-resolution theory and summarize key results from DFSQ [4], [5].…”