The traditional image rendition process assumes an ideal output imaging system where each image pixel will be perfectly reproduced regardless of different image formation processes. In reality, each individual imaging system can be approximated as a linear system with its own unique transfer function characteristics, which could exhibit spatial and/or temporal variations. Consequently, it will produce inconsistent tone scale reproduction curves during the print production run. The objective of this paper is to investigate the tone reproduction curve variations caused primarily by the interaction of neighboring imaging pixels. An intelligent image rendition algorithm is presented to compensate for this macroscopic imaging process deficiency. Experimental results will be demonstrated to show the effectiveness of the proposed algorithm, which has been deployed commercially on a digital printing press.