Use this URL to cite or link to this record in EThOS:
Title: Verification of colour appearance models using magnitude estimation data
Author: Juan, Lu-Yin Grace
ISNI:       0000 0001 3593 3968
Awarding Body: University of Derby
Current Institution: University of Derby
Date of Award: 2000
Availability of Full Text:
Access from EThOS:
Access from Institution:
A reliable colour appearance model is highly desired by various industries such as textiles, paint, plastic, colour imaging, etc. Five colour appearance models named CAM97s3, CAM97s3c, CAM97s4, CAM97s4c and CAM97s5, which are capable of accurately predicting the colour appearance under different viewing conditions, were developed from this study. They were modified from the current international standard colour appearancem odel CIIECAM97s. A series of extensive psychophysical experiments was conducted to scale the colour appearanceo f object colours using magnitudee stimation methodsu nder twelve phaseso f various viewing conditions in this study. The viewing parameters of the samples included two sizes (large and small), three backgrounds (white, grey and black), two textures (paint and textile) and from perceived colour attributes (lightness, colourfulness, hue and saturation). In total, 40,254 estimations were made. The purposes of this project were: 1. to investigate the influence of the size, background and texture on the perception of three different colour attributes - the lightness, colourfulness and hue; 2. to examine the Helmholtz-Kohlrausch (HK) effect, i. e. the lightness affected by chroma of a sample; 3. to accumulate the saturation experimental data; 4. to test the performanceo f the available colour appearancem odels;a nd 5. to improve existing functions in terms of accuracy and simplicity, and to develop a new saturation scale for the CIIECAM97s type models. This thesisd escribesth e experimentaml ethodsc, ollectiono f the psychophysicadl.a ta, analysis of observer performance and various visual effects. A unique data set in visual experiments was collected and is named the Juan & Luo data set. The results were analyzed and discussed. It, was found that the lightness contrast effect is occurred for both large and small sizes. In general, colours appear darker against a lighter background. The results show that there is no size effect for all three attributes studied: lightness, colourfulness and hue. The textile samples appear slightly lighter and more colourful than the paint samples for all three neutral backgrounds. This study confirmed that the HK effect makes most chromatic colours appear lighter than the neutral colours which have similar luminance. The results also indicate that observers can be trained to scale saturation with a great accuracy. By dividing colourfulness to lightness visual results, a strong correlation between these values and saturation results can be found. Twelve colour appearancem odels were tested using the Juan & Luo results. They can be divided into seven different types: CIELAB, Nayatani, RLAB, Hunt, LLAB, ATD, CIECAM97s. (Each type may have more than one version.) Comparing lightness predictions from different models, ATD and RLAB models slightly outperformed the other models. The CIECAM type models gave the most accurate predictions to the visual colourfulness results amongst all the models tested. All models perform the same good in hue prediction and all gave poor predictions to the data of saturation. Efforts were made to include not just the present but also previously accumulated data sets to derive or modify new scales to improve the performance of the CIECAM97s colour appearancem odel. Finally, five models were achieved:C AM97s3, CAM97s3c, CAM97s4, CAM97s4c and CAM97s5. The former two are more like CIECAM97s, the latter are a simplification of lightness, brightness, chroma, colourfulness and saturation scales to the CIECAM97s. All new models perform nearly the same but CAM97s5 is the best.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: Visual; Juan; Luo; Colourfulness