Existing colorization algorithms can hardly provide pleasant visual effect in hi ghlight or shadow areas.Motivated by the general imaging principle of natural co lor images,a realism improvement approach based on post-process correction in c hrominance is proposed.First,an imaging model is built,where the influence of bo th light and objective on resulting image is decoupled.Then,a further chrominanc e-luminance relation model for any pixel imaged in white light is established a nd employed to identify the very reason for falsehood resulting from existing co lorization methods,and a realism improvement strategy is proposed based on chrom inance coorection.Since only a grey parameter is to be set,two convenient operat ion modes are available—automatic coarse adjustment based on default values,and fine tuning by user′s intervention.In addition,this decupling model makes it e asy to extend colorization to case of color light.Experiment results indicate th at the visual effect after chrominance correction provides more accurate feature significantly in highlight or shadow areas and even outperforms that of existin g algorithms under expensive and time-consuming miscellaneous input.The runtime of every correction is less than 0.3s,which means this operation could perform as human-computer interaction in real-time as well,and hence demonstrates tha t this proposed method could provide a general platform to improve realism of co lorization result.