RGB Approach for Pixel-Wise Identification of Cellulose Nitrate Photo Negative Yellowing

Abstract

Film-based cellulose nitrate negatives are a unique class of objects that contain important information about life, historical buildings, and the natural landscapes of past years. Increased sensitivity to storage conditions makes these objects highly flammable and can lead to irretrievable loss. In this regard, timely identification of the degradation process is a necessary step towards further conservation and restoration. This work studies the possibility of detecting the degradation process based on cellulose nitrate artifact yellowing. A total of 20 normal and 20 yellowed negatives from the collection of Karl Kosse (The State Museum and Exhibition Center ROSPHOTO) were selected as objects for statistical study. The novelty of this work is in its demonstration of the possibility to divide negatives into normal and yellowed areas with different shades based on different B/R and B/G ratios of both light and dark negatives, i.e., regardless of the distribution of RGB component values for the obtained digital photo from the negative. Moreover, the obtained differentiation result was demonstrated for individual image pixels, without the need for averaging over a certain area.

Keywords

• cellulose nitrate
• RGB components
• statistics
• yellowing