An evaluation of quality metrics for compressed images based on human visual sensitivity

摘要: It is generally believed that the performance of current image coding techniques not close enough to fundamental limit on bit rate at which high reception quality maintained. For supporting future applications demand very low rates, more efficient algorithms are expected. Among numerous approaches reach this optimality, perceptual matches compression algorithm human perception mechanism has been considered as most promising solution, and recently become an important area research. In addition investigation coding, need for a distortion metric predicting long recognized in processing community. still evaluation, several measures tied some psycho-visual properties have proposed. Examples such be found literature. One significant problem associated with these metrics there little information how perform comparison one another. The purpose paper provide rigorous evaluation four designed coded achromatic images.