A Study on Subjective Evaluation of Printed Launch Vehicle Lift-Off View Colour Photographs and Pattern Quantification of Perceived Image Quality Attributes.

Quantification of subjectivity is essential to have a consistent quality in prints for further analysis of rocket lift-off view printed images. An experimental mathematical model has been derived to quantify the human visual response vis a vis print process controls in rocket lift-off view sequence photography. Subjective evaluation data of eight professional printing technicians was collected using pairwise comparison. They gave 16 kinds of evaluations for each print. Their data consistency was checked by the process clustering of the subjects and qualified for taking as a quantification model for the analysis of perception in print quality. After trying multi variant analysis and judgmental analysis, it was found that the problem is fuzzy in nature and involves multiple linkages and hence human like response is required. Hence, neural network model was selected which works on a simplistic non-linear weighted averages attached to inputs projected from dynamic modification of the target to reach the target output. The interdependency in the factor analysis output is linked up in a feed forward neural network. The 16th attribute of the data evaluation (overall preference) parameter is taken as output. The inter linkages of the possible explanation from the contributors of the remaining parameters is explained by the performance of the neural network. The positive correlation of the factors confirms the data with that of the nationally acclaimed professional colour printing laboratory expert’s opinion.