In this study, an affective sensing approach is proposed to evaluate the computer user's affective states. This is achieved by the processing of the pupil diameter (PD) signal from an eye gaze tracker and its classification with a support vector machine (SVM). Previous research has found that the diameter of the human pupil can be seen as an indication of affective processing. However, the pupillary light reflex (PLR), which decreases the pupil size as larger amounts of light are sensed by the retina, is known as the dominant factor in resizing the pupil. Therefore, an adaptive interference canceller (AIC) system using the H-infinity time-varying (HITV) adaptive algorithm was developed to minimize the impact of the PLR on the measured PD signal. Under that assumption, the output of the AIC, i.e., the modified pupil diameter (MPD) signal, was used to identify the pupillary affective responses (PAR) of the subject. Additional manipulations of the MPD signal yield the Processed MPD (PMPD) signal, and its mean is used as a feature for the recognition of affective states (e.g., stress) by means of a support vector machine (SVM) classifier. The efficiency of affective sensing through the PD signal was evaluated by comparison to the corresponding SVM classifications through features derived from the simultaneously recorded galvanic skin response (GSR) and blood volume pulse (BVP) signals, with and without the PD feature. This study confirmed benefits of using an AIC with the HITV adaptive algorithm to minimize the PD changes caused by PLR (to an acceptable level), and therefore facilitate the affective sensing of a computer user through the PD signal processing. Copyright 2010 ISA. All Rights Reserved.