In digital environments, colour is used for many purposes: for example, to encode information in charts, signify missing field information on websites, and identify active windows and menus. However, many people have inherited, acquired, or situationally-induced Colour Vision Deficiency (CVD), and therefore have difficulties differentiating many colours. Recolouring tools have been developed that modify interface colours to make them more differentiable for people with CVD, but these tools rely on models of colour differentiation that do not represent the majority of people with CVD. As a result, existing recolouring tools do not help most people with CVD. To solve this problem, I developed Situation-Specific Modelling (SSM), and applied it to colour differentiation to develop the Individualized model of Colour Differentiation (ICD). SSM utilizes an in-situ calibration procedure to measure a particular user’s abilities within a particular situation, and a modelling component to extend the calibration measurements into a full representation of the user’s abilities. ICD applies in-situ calibration to measuring a user’s unique colour differentiation abilities, and contains a modelling component that is capable of representing the colour differentiation abilities of almost any individual with CVD. This dissertation presents four versions of the ICD and one application of the ICD to recolouring. First, I describe the development and evaluation of a feasibility implementation of the ICD that tests the viability of the SSM approach. Second, I present revised calibration and modelling components of the ICD that reduce the calibration time from 32 minutes to two minutes. Next, I describe the third and fourth ICD versions that improve the applicability of the ICD to recolouring tools by reducing the colour differentiation prediction time and increasing the power of each prediction. Finally, I present a new recolouring tool (ICDRecolour) that uses the ICD model to steer the recolouring process. In a comparative evaluation, ICDRecolour achieved 90% colour matching accuracy for participants – 20% better than existing recolouring tools – for a wide range of CVDs. By modelling the colour differentiation abilities of a particular user in a particular environment, the ICD enables the extension of recolouring tools to helping most people with CVD, thereby reducing the difficulties that people with CVD experience when using colour in digital environments.