Individualised Halo-Free Gradient-Domain Colour Image Daltonisation

Area of Science:

• Computer Vision
• Image Processing
• Human-Computer Interaction

Background:

• Daltonisation aims to restore visibility of image details for colour vision deficient (CVD) observers.
• Previous methods often introduce artificial colours and may produce artefacts.
• A prior gradient-domain method showed promising results in behavioural and psychometric tests.

Purpose of the Study:

• To improve a gradient-domain colour image daltonisation method.
• To reduce computational time and eliminate halo artefacts.
• To enhance accessibility of visual information for individuals with CVD.

Main Methods:

• Improved initial estimation for faster daltonisation.
• Incorporated local linear anisotropic diffusion to remove artefacts.
• Utilised a colour vision deficiency simulation algorithm for broad applicability.

Main Results:

• Significantly reduced computational time for daltonisation.
• Effectively removed halo artefacts, improving image quality.
• Demonstrated applicability to various types of CVD and individualised vision.

Conclusions:

• The enhanced daltonisation method offers improved efficiency and image quality.
• This advancement increases the accessibility of visual content for CVD observers.
• The method's adaptability allows for personalised visual restoration.