Image Compression Based on Mean Value Predictive Vector Quantization

The main drawback of traditional predictive vector quantization (PVQ) is the time consuming encoding process. In order to reduce the computational complexity, an algorithm called mean value predictive vector quantization (MVPVQ) is proposed in this paper. Input vectors are classified into smooth vectors and non-smooth vectors before encoding. Different kinds of input vectors are encoded with different encoding schemes. Simulations demonstrate that the proposed method can achieve satisfying image quality at low bit rates.